public inbox for [email protected]
help / color / mirror / Atom feed[PATCH v1] Fix usage of unified logging pg_log_* in pg_rewind and initdb
166+ messages / 13 participants
[nested] [flat]
* [PATCH v1] Fix usage of unified logging pg_log_* in pg_rewind and initdb
@ 2019-07-01 15:11 Alexey Kondratov <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Alexey Kondratov @ 2019-07-01 15:11 UTC (permalink / raw)
---
src/bin/initdb/initdb.c | 2 +-
src/bin/pg_rewind/pg_rewind.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/bin/initdb/initdb.c b/src/bin/initdb/initdb.c
index 2ef179165b..70273be783 100644
--- a/src/bin/initdb/initdb.c
+++ b/src/bin/initdb/initdb.c
@@ -2497,7 +2497,7 @@ setup_bin_paths(const char *argv0)
pg_log_error("The program \"postgres\" is needed by %s but was not found in the\n"
"same directory as \"%s\".\n"
"Check your installation.",
- full_path, progname);
+ progname, full_path);
else
pg_log_error("The program \"postgres\" was found by \"%s\"\n"
"but was not the same version as %s.\n"
diff --git a/src/bin/pg_rewind/pg_rewind.c b/src/bin/pg_rewind/pg_rewind.c
index 6e77201be6..d378053de4 100644
--- a/src/bin/pg_rewind/pg_rewind.c
+++ b/src/bin/pg_rewind/pg_rewind.c
@@ -555,7 +555,7 @@ getTimelineHistory(ControlFileData *controlFile, int *nentries)
else if (controlFile == &ControlFile_target)
histfile = slurpFile(datadir_target, path, NULL);
else
- pg_fatal("invalid control file\n");
+ pg_fatal("invalid control file");
history = rewind_parseTimeLineHistory(histfile, tli, nentries);
pg_free(histfile);
base-commit: 95bbe5d82e428db342fa3ec60b95f1b9873741e5
--
2.17.1
--------------0DB63CE8F653DAD250E3D5A4--
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH 1/2] Unique key
@ 2019-11-11 13:49 jesperpedersen <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: jesperpedersen @ 2019-11-11 13:49 UTC (permalink / raw)
Design by David Rowley.
Author: Jesper Pedersen
---
src/backend/nodes/outfuncs.c | 14 +++
src/backend/nodes/print.c | 39 +++++++
src/backend/optimizer/path/Makefile | 3 +-
src/backend/optimizer/path/allpaths.c | 8 ++
src/backend/optimizer/path/costsize.c | 5 +
src/backend/optimizer/path/indxpath.c | 41 +++++++
src/backend/optimizer/path/pathkeys.c | 71 ++++++++++--
src/backend/optimizer/path/uniquekey.c | 147 +++++++++++++++++++++++++
src/backend/optimizer/plan/planagg.c | 1 +
src/backend/optimizer/plan/planmain.c | 1 +
src/backend/optimizer/plan/planner.c | 17 ++-
src/backend/optimizer/util/pathnode.c | 12 ++
src/include/nodes/nodes.h | 1 +
src/include/nodes/pathnodes.h | 18 +++
src/include/nodes/print.h | 1 +
src/include/optimizer/pathnode.h | 1 +
src/include/optimizer/paths.h | 11 ++
17 files changed, 378 insertions(+), 13 deletions(-)
create mode 100644 src/backend/optimizer/path/uniquekey.c
diff --git a/src/backend/nodes/outfuncs.c b/src/backend/nodes/outfuncs.c
index b0dcd02ff6..1ccd68d3aa 100644
--- a/src/backend/nodes/outfuncs.c
+++ b/src/backend/nodes/outfuncs.c
@@ -1720,6 +1720,7 @@ _outPathInfo(StringInfo str, const Path *node)
=09WRITE_FLOAT_FIELD(startup_cost, "%.2f");
=09WRITE_FLOAT_FIELD(total_cost, "%.2f");
=09WRITE_NODE_FIELD(pathkeys);
+=09WRITE_NODE_FIELD(uniquekeys);
}
=20
/*
@@ -2201,6 +2202,7 @@ _outPlannerInfo(StringInfo str, const PlannerInfo *no=
de)
=09WRITE_NODE_FIELD(eq_classes);
=09WRITE_BOOL_FIELD(ec_merging_done);
=09WRITE_NODE_FIELD(canon_pathkeys);
+=09WRITE_NODE_FIELD(canon_uniquekeys);
=09WRITE_NODE_FIELD(left_join_clauses);
=09WRITE_NODE_FIELD(right_join_clauses);
=09WRITE_NODE_FIELD(full_join_clauses);
@@ -2210,6 +2212,7 @@ _outPlannerInfo(StringInfo str, const PlannerInfo *no=
de)
=09WRITE_NODE_FIELD(placeholder_list);
=09WRITE_NODE_FIELD(fkey_list);
=09WRITE_NODE_FIELD(query_pathkeys);
+=09WRITE_NODE_FIELD(query_uniquekeys);
=09WRITE_NODE_FIELD(group_pathkeys);
=09WRITE_NODE_FIELD(window_pathkeys);
=09WRITE_NODE_FIELD(distinct_pathkeys);
@@ -2397,6 +2400,14 @@ _outPathKey(StringInfo str, const PathKey *node)
=09WRITE_BOOL_FIELD(pk_nulls_first);
}
=20
+static void
+_outUniqueKey(StringInfo str, const UniqueKey *node)
+{
+=09WRITE_NODE_TYPE("UNIQUEKEY");
+
+=09WRITE_NODE_FIELD(eq_clause);
+}
+
static void
_outPathTarget(StringInfo str, const PathTarget *node)
{
@@ -4083,6 +4094,9 @@ outNode(StringInfo str, const void *obj)
=09=09=09case T_PathKey:
=09=09=09=09_outPathKey(str, obj);
=09=09=09=09break;
+=09=09=09case T_UniqueKey:
+=09=09=09=09_outUniqueKey(str, obj);
+=09=09=09=09break;
=09=09=09case T_PathTarget:
=09=09=09=09_outPathTarget(str, obj);
=09=09=09=09break;
diff --git a/src/backend/nodes/print.c b/src/backend/nodes/print.c
index 4ecde6b421..62c9d4ef7a 100644
--- a/src/backend/nodes/print.c
+++ b/src/backend/nodes/print.c
@@ -459,6 +459,45 @@ print_pathkeys(const List *pathkeys, const List *rtabl=
e)
=09printf(")\n");
}
=20
+/*
+ * print_uniquekeys -
+ *=09 unique_key an UniqueKey
+ */
+void
+print_uniquekeys(const List *uniquekeys, const List *rtable)
+{
+=09ListCell *l;
+
+=09printf("(");
+=09foreach(l, uniquekeys)
+=09{
+=09=09UniqueKey *unique_key =3D (UniqueKey *) lfirst(l);
+=09=09EquivalenceClass *eclass =3D (EquivalenceClass *) unique_key->eq_cla=
use;
+=09=09ListCell *k;
+=09=09bool=09=09first =3D true;
+
+=09=09/* chase up */
+=09=09while (eclass->ec_merged)
+=09=09=09eclass =3D eclass->ec_merged;
+
+=09=09printf("(");
+=09=09foreach(k, eclass->ec_members)
+=09=09{
+=09=09=09EquivalenceMember *mem =3D (EquivalenceMember *) lfirst(k);
+
+=09=09=09if (first)
+=09=09=09=09first =3D false;
+=09=09=09else
+=09=09=09=09printf(", ");
+=09=09=09print_expr((Node *) mem->em_expr, rtable);
+=09=09}
+=09=09printf(")");
+=09=09if (lnext(uniquekeys, l))
+=09=09=09printf(", ");
+=09}
+=09printf(")\n");
+}
+
/*
* print_tl
*=09 print targetlist in a more legible way.
diff --git a/src/backend/optimizer/path/Makefile b/src/backend/optimizer/pa=
th/Makefile
index 1e199ff66f..63cc1505d9 100644
--- a/src/backend/optimizer/path/Makefile
+++ b/src/backend/optimizer/path/Makefile
@@ -21,6 +21,7 @@ OBJS =3D \
=09joinpath.o \
=09joinrels.o \
=09pathkeys.o \
-=09tidpath.o
+=09tidpath.o \
+=09uniquekey.o
=20
include $(top_srcdir)/src/backend/common.mk
diff --git a/src/backend/optimizer/path/allpaths.c b/src/backend/optimizer/=
path/allpaths.c
index db3a68a51d..5fc9b81746 100644
--- a/src/backend/optimizer/path/allpaths.c
+++ b/src/backend/optimizer/path/allpaths.c
@@ -3954,6 +3954,14 @@ print_path(PlannerInfo *root, Path *path, int indent=
)
=09=09print_pathkeys(path->pathkeys, root->parse->rtable);
=09}
=20
+=09if (path->uniquekeys)
+=09{
+=09=09for (i =3D 0; i < indent; i++)
+=09=09=09printf("\t");
+=09=09printf(" uniquekeys: ");
+=09=09print_uniquekeys(path->uniquekeys, root->parse->rtable);
+=09}
+
=09if (join)
=09{
=09=09JoinPath *jp =3D (JoinPath *) path;
diff --git a/src/backend/optimizer/path/costsize.c b/src/backend/optimizer/=
path/costsize.c
index c5f6593485..0ec9a6db76 100644
--- a/src/backend/optimizer/path/costsize.c
+++ b/src/backend/optimizer/path/costsize.c
@@ -705,6 +705,11 @@ cost_index(IndexPath *path, PlannerInfo *root, double =
loop_count,
=09=09path->path.parallel_aware =3D true;
=09}
=20
+=09/* Consider cost based on unique key */
+=09if (path->path.uniquekeys)
+=09{
+=09}
+
=09/*
=09 * Now interpolate based on estimated index order correlation to get to=
tal
=09 * disk I/O cost for main table accesses.
diff --git a/src/backend/optimizer/path/indxpath.c b/src/backend/optimizer/=
path/indxpath.c
index 37b257cd0e..aa0da49119 100644
--- a/src/backend/optimizer/path/indxpath.c
+++ b/src/backend/optimizer/path/indxpath.c
@@ -189,6 +189,7 @@ static Expr *match_clause_to_ordering_op(IndexOptInfo *=
index,
static bool ec_member_matches_indexcol(PlannerInfo *root, RelOptInfo *rel,
=09=09=09=09=09=09=09=09=09 EquivalenceClass *ec, EquivalenceMember *em,
=09=09=09=09=09=09=09=09=09 void *arg);
+static List *get_uniquekeys_for_index(PlannerInfo *root, List *pathkeys);
=20
=20
/*
@@ -874,6 +875,7 @@ build_index_paths(PlannerInfo *root, RelOptInfo *rel,
=09List=09 *orderbyclausecols;
=09List=09 *index_pathkeys;
=09List=09 *useful_pathkeys;
+=09List=09 *useful_uniquekeys;
=09bool=09=09found_lower_saop_clause;
=09bool=09=09pathkeys_possibly_useful;
=09bool=09=09index_is_ordered;
@@ -1036,11 +1038,15 @@ build_index_paths(PlannerInfo *root, RelOptInfo *re=
l,
=09if (index_clauses !=3D NIL || useful_pathkeys !=3D NIL || useful_predic=
ate ||
=09=09index_only_scan)
=09{
+=09=09if (has_useful_uniquekeys(root))
+=09=09=09useful_uniquekeys =3D get_uniquekeys_for_index(root, useful_pathk=
eys);
+
=09=09ipath =3D create_index_path(root, index,
=09=09=09=09=09=09=09=09 index_clauses,
=09=09=09=09=09=09=09=09 orderbyclauses,
=09=09=09=09=09=09=09=09 orderbyclausecols,
=09=09=09=09=09=09=09=09 useful_pathkeys,
+=09=09=09=09=09=09=09=09 useful_uniquekeys,
=09=09=09=09=09=09=09=09 index_is_ordered ?
=09=09=09=09=09=09=09=09 ForwardScanDirection :
=09=09=09=09=09=09=09=09 NoMovementScanDirection,
@@ -1063,6 +1069,7 @@ build_index_paths(PlannerInfo *root, RelOptInfo *rel,
=09=09=09=09=09=09=09=09=09 orderbyclauses,
=09=09=09=09=09=09=09=09=09 orderbyclausecols,
=09=09=09=09=09=09=09=09=09 useful_pathkeys,
+=09=09=09=09=09=09=09=09=09 useful_uniquekeys,
=09=09=09=09=09=09=09=09=09 index_is_ordered ?
=09=09=09=09=09=09=09=09=09 ForwardScanDirection :
=09=09=09=09=09=09=09=09=09 NoMovementScanDirection,
@@ -1093,11 +1100,15 @@ build_index_paths(PlannerInfo *root, RelOptInfo *re=
l,
=09=09=09=09=09=09=09=09=09=09=09=09=09index_pathkeys);
=09=09if (useful_pathkeys !=3D NIL)
=09=09{
+=09=09=09if (has_useful_uniquekeys(root))
+=09=09=09=09useful_uniquekeys =3D get_uniquekeys_for_index(root, useful_pa=
thkeys);
+
=09=09=09ipath =3D create_index_path(root, index,
=09=09=09=09=09=09=09=09=09 index_clauses,
=09=09=09=09=09=09=09=09=09 NIL,
=09=09=09=09=09=09=09=09=09 NIL,
=09=09=09=09=09=09=09=09=09 useful_pathkeys,
+=09=09=09=09=09=09=09=09=09 useful_uniquekeys,
=09=09=09=09=09=09=09=09=09 BackwardScanDirection,
=09=09=09=09=09=09=09=09=09 index_only_scan,
=09=09=09=09=09=09=09=09=09 outer_relids,
@@ -1115,6 +1126,7 @@ build_index_paths(PlannerInfo *root, RelOptInfo *rel,
=09=09=09=09=09=09=09=09=09=09 NIL,
=09=09=09=09=09=09=09=09=09=09 NIL,
=09=09=09=09=09=09=09=09=09=09 useful_pathkeys,
+=09=09=09=09=09=09=09=09=09=09 useful_uniquekeys,
=09=09=09=09=09=09=09=09=09=09 BackwardScanDirection,
=09=09=09=09=09=09=09=09=09=09 index_only_scan,
=09=09=09=09=09=09=09=09=09=09 outer_relids,
@@ -3365,6 +3377,35 @@ match_clause_to_ordering_op(IndexOptInfo *index,
=09return clause;
}
=20
+/*
+ * get_uniquekeys_for_index
+ */
+static List *
+get_uniquekeys_for_index(PlannerInfo *root, List *pathkeys)
+{
+=09ListCell *lc;
+
+=09if (pathkeys)
+=09{
+=09=09List *uniquekeys =3D NIL;
+=09=09foreach(lc, pathkeys)
+=09=09{
+=09=09=09UniqueKey *unique_key;
+=09=09=09PathKey *pk =3D (PathKey *) lfirst(lc);
+=09=09=09EquivalenceClass *ec =3D (EquivalenceClass *) pk->pk_eclass;
+
+=09=09=09unique_key =3D makeNode(UniqueKey);
+=09=09=09unique_key->eq_clause =3D ec;
+=09=09=09
+=09=09=09lappend(uniquekeys, unique_key);
+=09=09}
+
+=09=09if (uniquekeys_contained_in(root->canon_uniquekeys, uniquekeys))
+=09=09=09return uniquekeys;
+=09}
+
+=09return NIL;
+}
=20
/*************************************************************************=
***
*=09=09=09=09---- ROUTINES TO DO PARTIAL INDEX PREDICATE TESTS=09----
diff --git a/src/backend/optimizer/path/pathkeys.c b/src/backend/optimizer/=
path/pathkeys.c
index 2f4fea241a..b0dc1bc22a 100644
--- a/src/backend/optimizer/path/pathkeys.c
+++ b/src/backend/optimizer/path/pathkeys.c
@@ -29,6 +29,7 @@
#include "utils/lsyscache.h"
=20
=20
+static bool pathkey_is_unique(PathKey *new_pathkey, List *pathkeys);
static bool pathkey_is_redundant(PathKey *new_pathkey, List *pathkeys);
static bool matches_boolean_partition_clause(RestrictInfo *rinfo,
=09=09=09=09=09=09=09=09=09=09=09 RelOptInfo *partrel,
@@ -96,6 +97,29 @@ make_canonical_pathkey(PlannerInfo *root,
=09return pk;
}
=20
+/*
+ * pathkey_is_unique
+ *=09 Checks if the new pathkey's equivalence class is the same as that =
of
+ * any existing member of the pathkey list.
+ */
+static bool
+pathkey_is_unique(PathKey *new_pathkey, List *pathkeys)
+{
+=09EquivalenceClass *new_ec =3D new_pathkey->pk_eclass;
+=09ListCell *lc;
+
+=09/* If same EC already is already in the list, then not unique */
+=09foreach(lc, pathkeys)
+=09{
+=09=09PathKey *old_pathkey =3D (PathKey *) lfirst(lc);
+
+=09=09if (new_ec =3D=3D old_pathkey->pk_eclass)
+=09=09=09return false;
+=09}
+
+=09return true;
+}
+
/*
* pathkey_is_redundant
*=09 Is a pathkey redundant with one already in the given list?
@@ -135,22 +159,12 @@ static bool
pathkey_is_redundant(PathKey *new_pathkey, List *pathkeys)
{
=09EquivalenceClass *new_ec =3D new_pathkey->pk_eclass;
-=09ListCell *lc;
=20
=09/* Check for EC containing a constant --- unconditionally redundant */
=09if (EC_MUST_BE_REDUNDANT(new_ec))
=09=09return true;
=20
-=09/* If same EC already used in list, then redundant */
-=09foreach(lc, pathkeys)
-=09{
-=09=09PathKey *old_pathkey =3D (PathKey *) lfirst(lc);
-
-=09=09if (new_ec =3D=3D old_pathkey->pk_eclass)
-=09=09=09return true;
-=09}
-
-=09return false;
+=09return !pathkey_is_unique(new_pathkey, pathkeys);
}
=20
/*
@@ -1098,6 +1112,41 @@ make_pathkeys_for_sortclauses(PlannerInfo *root,
=09return pathkeys;
}
=20
+/*
+ * make_pathkeys_for_uniquekeyclauses
+ *=09=09Generate a pathkeys list to be used for uniquekey clauses
+ */
+List *
+make_pathkeys_for_uniquekeys(PlannerInfo *root,
+=09=09=09=09=09=09=09 List *sortclauses,
+=09=09=09=09=09=09=09 List *tlist)
+{
+=09List=09 *pathkeys =3D NIL;
+=09ListCell *l;
+
+=09foreach(l, sortclauses)
+=09{
+=09=09SortGroupClause *sortcl =3D (SortGroupClause *) lfirst(l);
+=09=09Expr=09 *sortkey;
+=09=09PathKey *pathkey;
+
+=09=09sortkey =3D (Expr *) get_sortgroupclause_expr(sortcl, tlist);
+=09=09Assert(OidIsValid(sortcl->sortop));
+=09=09pathkey =3D make_pathkey_from_sortop(root,
+=09=09=09=09=09=09=09=09=09=09 sortkey,
+=09=09=09=09=09=09=09=09=09=09 root->nullable_baserels,
+=09=09=09=09=09=09=09=09=09=09 sortcl->sortop,
+=09=09=09=09=09=09=09=09=09=09 sortcl->nulls_first,
+=09=09=09=09=09=09=09=09=09=09 sortcl->tleSortGroupRef,
+=09=09=09=09=09=09=09=09=09=09 true);
+
+=09=09if (pathkey_is_unique(pathkey, pathkeys))
+=09=09=09pathkeys =3D lappend(pathkeys, pathkey);
+=09}
+
+=09return pathkeys;
+}
+
/*************************************************************************=
***
*=09=09PATHKEYS AND MERGECLAUSES
*************************************************************************=
***/
diff --git a/src/backend/optimizer/path/uniquekey.c b/src/backend/optimizer=
/path/uniquekey.c
new file mode 100644
index 0000000000..13d4ebb98c
--- /dev/null
+++ b/src/backend/optimizer/path/uniquekey.c
@@ -0,0 +1,147 @@
+/*------------------------------------------------------------------------=
-
+ *
+ * uniquekey.c
+ *=09 Utilities for matching and building unique keys
+ *
+ * Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group
+ * Portions Copyright (c) 1994, Regents of the University of California
+ *
+ * IDENTIFICATION
+ *=09 src/backend/optimizer/path/uniquekey.c
+ *
+ *------------------------------------------------------------------------=
-
+ */
+#include "postgres.h"
+
+#include "optimizer/pathnode.h"
+#include "optimizer/paths.h"
+#include "nodes/pg_list.h"
+
+static UniqueKey *make_canonical_uniquekey(PlannerInfo *root, EquivalenceC=
lass *eclass);
+
+/*
+ * Build a list of unique keys
+ */
+List*
+build_uniquekeys(PlannerInfo *root, List *sortclauses)
+{
+=09List *result =3D NIL;
+=09List *sortkeys;
+=09ListCell *l;
+
+=09sortkeys =3D make_pathkeys_for_uniquekeys(root,
+=09=09=09=09=09=09=09=09=09=09=09sortclauses,
+=09=09=09=09=09=09=09=09=09=09=09root->processed_tlist);
+
+=09/* Create a uniquekey and add it to the list */
+=09foreach(l, sortkeys)
+=09{
+=09=09PathKey *pathkey =3D (PathKey *) lfirst(l);
+=09=09EquivalenceClass *ec =3D pathkey->pk_eclass;
+=09=09UniqueKey *unique_key =3D make_canonical_uniquekey(root, ec);
+
+=09=09result =3D lappend(result, unique_key);
+=09}
+
+=09return result;
+}
+
+/*
+ * uniquekeys_contained_in
+ *=09 Are the keys2 included in the keys1 superset
+ */
+bool
+uniquekeys_contained_in(List *keys1, List *keys2)
+{
+=09ListCell *key1,
+=09=09=09 *key2;
+
+=09/*
+=09 * Fall out quickly if we are passed two identical lists. This mostly
+=09 * catches the case where both are NIL, but that's common enough to
+=09 * warrant the test.
+=09 */
+=09if (keys1 =3D=3D keys2)
+=09=09return true;
+
+=09foreach(key2, keys2)
+=09{
+=09=09bool found =3D false;
+=09=09UniqueKey *uniquekey2 =3D (UniqueKey *) lfirst(key2);
+
+=09=09foreach(key1, keys1)
+=09=09{
+=09=09=09UniqueKey *uniquekey1 =3D (UniqueKey *) lfirst(key1);
+
+=09=09=09if (uniquekey1->eq_clause =3D=3D uniquekey2->eq_clause)
+=09=09=09{
+=09=09=09=09found =3D true;
+=09=09=09=09break;
+=09=09=09}
+=09=09}
+
+=09=09if (!found)
+=09=09=09return false;
+=09}
+
+=09return true;
+}
+
+/*
+ * has_useful_uniquekeys
+ *=09=09Detect whether the planner could have any uniquekeys that are
+ *=09=09useful.
+ */
+bool
+has_useful_uniquekeys(PlannerInfo *root)
+{
+=09if (root->query_uniquekeys !=3D NIL)
+=09=09return true;=09/* there are some */
+=09return false;=09=09/* definitely useless */
+}
+
+/*
+ * make_canonical_uniquekey
+ *=09 Given the parameters for a UniqueKey, find any pre-existing matchin=
g
+ *=09 uniquekey in the query's list of "canonical" uniquekeys. Make a ne=
w
+ *=09 entry if there's not one already.
+ *
+ * Note that this function must not be used until after we have completed
+ * merging EquivalenceClasses. (We don't try to enforce that here; instea=
d,
+ * equivclass.c will complain if a merge occurs after root->canon_uniqueke=
ys
+ * has become nonempty.)
+ */
+static UniqueKey *
+make_canonical_uniquekey(PlannerInfo *root,
+=09=09=09=09=09=09 EquivalenceClass *eclass)
+{
+=09UniqueKey *uk;
+=09ListCell *lc;
+=09MemoryContext oldcontext;
+
+=09/* The passed eclass might be non-canonical, so chase up to the top */
+=09while (eclass->ec_merged)
+=09=09eclass =3D eclass->ec_merged;
+
+=09foreach(lc, root->canon_uniquekeys)
+=09{
+=09=09uk =3D (UniqueKey *) lfirst(lc);
+=09=09if (eclass =3D=3D uk->eq_clause)
+=09=09=09return uk;
+=09}
+
+=09/*
+=09 * Be sure canonical uniquekeys are allocated in the main planning cont=
ext.
+=09 * Not an issue in normal planning, but it is for GEQO.
+=09 */
+=09oldcontext =3D MemoryContextSwitchTo(root->planner_cxt);
+
+=09uk =3D makeNode(UniqueKey);
+=09uk->eq_clause =3D eclass;
+
+=09root->canon_uniquekeys =3D lappend(root->canon_uniquekeys, uk);
+
+=09MemoryContextSwitchTo(oldcontext);
+
+=09return uk;
+}
diff --git a/src/backend/optimizer/plan/planagg.c b/src/backend/optimizer/p=
lan/planagg.c
index 9381939c82..3d32f6dfd6 100644
--- a/src/backend/optimizer/plan/planagg.c
+++ b/src/backend/optimizer/plan/planagg.c
@@ -512,6 +512,7 @@ minmax_qp_callback(PlannerInfo *root, void *extra)
=09=09=09=09=09=09=09=09=09 root->parse->targetList);
=20
=09root->query_pathkeys =3D root->sort_pathkeys;
+=09root->query_uniquekeys =3D NIL;
}
=20
/*
diff --git a/src/backend/optimizer/plan/planmain.c b/src/backend/optimizer/=
plan/planmain.c
index f0c1b52a2e..3ccde03ab7 100644
--- a/src/backend/optimizer/plan/planmain.c
+++ b/src/backend/optimizer/plan/planmain.c
@@ -70,6 +70,7 @@ query_planner(PlannerInfo *root,
=09root->join_rel_level =3D NULL;
=09root->join_cur_level =3D 0;
=09root->canon_pathkeys =3D NIL;
+=09root->canon_uniquekeys =3D NIL;
=09root->left_join_clauses =3D NIL;
=09root->right_join_clauses =3D NIL;
=09root->full_join_clauses =3D NIL;
diff --git a/src/backend/optimizer/plan/planner.c b/src/backend/optimizer/p=
lan/planner.c
index 17c5f086fb..2507ec7d2a 100644
--- a/src/backend/optimizer/plan/planner.c
+++ b/src/backend/optimizer/plan/planner.c
@@ -3652,15 +3652,30 @@ standard_qp_callback(PlannerInfo *root, void *extra=
)
=09 * much easier, since we know that the parser ensured that one is a
=09 * superset of the other.
=09 */
+=09root->query_uniquekeys =3D NIL;
+
=09if (root->group_pathkeys)
+=09{
=09=09root->query_pathkeys =3D root->group_pathkeys;
+
+=09=09if (!root->parse->hasAggs)
+=09=09=09root->query_uniquekeys =3D build_uniquekeys(root, qp_extra->group=
Clause);
+=09}
=09else if (root->window_pathkeys)
=09=09root->query_pathkeys =3D root->window_pathkeys;
=09else if (list_length(root->distinct_pathkeys) >
=09=09=09 list_length(root->sort_pathkeys))
+=09{
=09=09root->query_pathkeys =3D root->distinct_pathkeys;
+=09=09root->query_uniquekeys =3D build_uniquekeys(root, parse->distinctCla=
use);
+=09}
=09else if (root->sort_pathkeys)
+=09{
=09=09root->query_pathkeys =3D root->sort_pathkeys;
+
+=09=09if (root->distinct_pathkeys)
+=09=09=09root->query_uniquekeys =3D build_uniquekeys(root, parse->distinct=
Clause);
+=09}
=09else
=09=09root->query_pathkeys =3D NIL;
}
@@ -6217,7 +6232,7 @@ plan_cluster_use_sort(Oid tableOid, Oid indexOid)
=20
=09/* Estimate the cost of index scan */
=09indexScanPath =3D create_index_path(root, indexInfo,
-=09=09=09=09=09=09=09=09=09 NIL, NIL, NIL, NIL,
+=09=09=09=09=09=09=09=09=09 NIL, NIL, NIL, NIL, NIL,
=09=09=09=09=09=09=09=09=09 ForwardScanDirection, false,
=09=09=09=09=09=09=09=09=09 NULL, 1.0, false);
=20
diff --git a/src/backend/optimizer/util/pathnode.c b/src/backend/optimizer/=
util/pathnode.c
index 34acb732ee..f268112386 100644
--- a/src/backend/optimizer/util/pathnode.c
+++ b/src/backend/optimizer/util/pathnode.c
@@ -941,6 +941,7 @@ create_seqscan_path(PlannerInfo *root, RelOptInfo *rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D parallel_workers;
=09pathnode->pathkeys =3D NIL;=09/* seqscan has unordered result */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_seqscan(pathnode, root, rel, pathnode->param_info);
=20
@@ -965,6 +966,7 @@ create_samplescan_path(PlannerInfo *root, RelOptInfo *r=
el, Relids required_outer
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* samplescan has unordered result */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_samplescan(pathnode, root, rel, pathnode->param_info);
=20
@@ -1001,6 +1003,7 @@ create_index_path(PlannerInfo *root,
=09=09=09=09 List *indexorderbys,
=09=09=09=09 List *indexorderbycols,
=09=09=09=09 List *pathkeys,
+=09=09=09=09 List *uniquekeys,
=09=09=09=09 ScanDirection indexscandir,
=09=09=09=09 bool indexonly,
=09=09=09=09 Relids required_outer,
@@ -1019,6 +1022,7 @@ create_index_path(PlannerInfo *root,
=09pathnode->path.parallel_safe =3D rel->consider_parallel;
=09pathnode->path.parallel_workers =3D 0;
=09pathnode->path.pathkeys =3D pathkeys;
+=09pathnode->path.uniquekeys =3D uniquekeys;
=20
=09pathnode->indexinfo =3D index;
=09pathnode->indexclauses =3D indexclauses;
@@ -1062,6 +1066,7 @@ create_bitmap_heap_path(PlannerInfo *root,
=09pathnode->path.parallel_safe =3D rel->consider_parallel;
=09pathnode->path.parallel_workers =3D parallel_degree;
=09pathnode->path.pathkeys =3D NIL;=09/* always unordered */
+=09pathnode->path.uniquekeys =3D NIL;
=20
=09pathnode->bitmapqual =3D bitmapqual;
=20
@@ -1923,6 +1928,7 @@ create_functionscan_path(PlannerInfo *root, RelOptInf=
o *rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D pathkeys;
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_functionscan(pathnode, root, rel, pathnode->param_info);
=20
@@ -1949,6 +1955,7 @@ create_tablefuncscan_path(PlannerInfo *root, RelOptIn=
fo *rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* result is always unordered */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_tablefuncscan(pathnode, root, rel, pathnode->param_info);
=20
@@ -1975,6 +1982,7 @@ create_valuesscan_path(PlannerInfo *root, RelOptInfo =
*rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* result is always unordered */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_valuesscan(pathnode, root, rel, pathnode->param_info);
=20
@@ -2000,6 +2008,7 @@ create_ctescan_path(PlannerInfo *root, RelOptInfo *re=
l, Relids required_outer)
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* XXX for now, result is always unordere=
d */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_ctescan(pathnode, root, rel, pathnode->param_info);
=20
@@ -2026,6 +2035,7 @@ create_namedtuplestorescan_path(PlannerInfo *root, Re=
lOptInfo *rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* result is always unordered */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_namedtuplestorescan(pathnode, root, rel, pathnode->param_info);
=20
@@ -2052,6 +2062,7 @@ create_resultscan_path(PlannerInfo *root, RelOptInfo =
*rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* result is always unordered */
+=09pathnode->uniquekeys =3D NIL;
=20
=09cost_resultscan(pathnode, root, rel, pathnode->param_info);
=20
@@ -2078,6 +2089,7 @@ create_worktablescan_path(PlannerInfo *root, RelOptIn=
fo *rel,
=09pathnode->parallel_safe =3D rel->consider_parallel;
=09pathnode->parallel_workers =3D 0;
=09pathnode->pathkeys =3D NIL;=09/* result is always unordered */
+=09pathnode->uniquekeys =3D NIL;
=20
=09/* Cost is the same as for a regular CTE scan */
=09cost_ctescan(pathnode, root, rel, pathnode->param_info);
diff --git a/src/include/nodes/nodes.h b/src/include/nodes/nodes.h
index bce2d59b0d..cbb6ba2586 100644
--- a/src/include/nodes/nodes.h
+++ b/src/include/nodes/nodes.h
@@ -261,6 +261,7 @@ typedef enum NodeTag
=09T_EquivalenceMember,
=09T_PathKey,
=09T_PathTarget,
+=09T_UniqueKey,
=09T_RestrictInfo,
=09T_IndexClause,
=09T_PlaceHolderVar,
diff --git a/src/include/nodes/pathnodes.h b/src/include/nodes/pathnodes.h
index 23a06d718e..6198c31cd4 100644
--- a/src/include/nodes/pathnodes.h
+++ b/src/include/nodes/pathnodes.h
@@ -267,6 +267,8 @@ struct PlannerInfo
=20
=09List=09 *canon_pathkeys; /* list of "canonical" PathKeys */
=20
+=09List=09 *canon_uniquekeys; /* list of "canonical" UniqueKeys */
+
=09List=09 *left_join_clauses;=09/* list of RestrictInfos for mergejoina=
ble
=09=09=09=09=09=09=09=09=09 * outer join clauses w/nonnullable var on
=09=09=09=09=09=09=09=09=09 * left */
@@ -295,6 +297,8 @@ struct PlannerInfo
=20
=09List=09 *query_pathkeys; /* desired pathkeys for query_planner() */
=20
+=09List=09 *query_uniquekeys; /* */
+
=09List=09 *group_pathkeys; /* groupClause pathkeys, if any */
=09List=09 *window_pathkeys;=09/* pathkeys of bottom window, if any */
=09List=09 *distinct_pathkeys;=09/* distinctClause pathkeys, if any */
@@ -1075,6 +1079,15 @@ typedef struct ParamPathInfo
=09List=09 *ppi_clauses;=09/* join clauses available from outer rels */
} ParamPathInfo;
=20
+/*
+ * UniqueKey
+ */
+typedef struct UniqueKey
+{
+=09NodeTag=09=09type;
+
+=09EquivalenceClass *eq_clause;=09/* equivalence class */
+} UniqueKey;
=20
/*
* Type "Path" is used as-is for sequential-scan paths, as well as some ot=
her
@@ -1104,6 +1117,9 @@ typedef struct ParamPathInfo
*
* "pathkeys" is a List of PathKey nodes (see above), describing the sort
* ordering of the path's output rows.
+ *
+ * "uniquekeys", if not NIL, is a list of UniqueKey nodes (see above),
+ * describing the XXX.
*/
typedef struct Path
{
@@ -1127,6 +1143,8 @@ typedef struct Path
=20
=09List=09 *pathkeys;=09=09/* sort ordering of path's output */
=09/* pathkeys is a List of PathKey nodes; see above */
+
+=09List=09 *uniquekeys;=09/* the unique keys, or NIL if none */
} Path;
=20
/* Macro for extracting a path's parameterization relids; beware double ev=
al */
diff --git a/src/include/nodes/print.h b/src/include/nodes/print.h
index cbff56a724..f1a7112877 100644
--- a/src/include/nodes/print.h
+++ b/src/include/nodes/print.h
@@ -28,6 +28,7 @@ extern char *pretty_format_node_dump(const char *dump);
extern void print_rt(const List *rtable);
extern void print_expr(const Node *expr, const List *rtable);
extern void print_pathkeys(const List *pathkeys, const List *rtable);
+extern void print_uniquekeys(const List *uniquekeys, const List *rtable);
extern void print_tl(const List *tlist, const List *rtable);
extern void print_slot(TupleTableSlot *slot);
=20
diff --git a/src/include/optimizer/pathnode.h b/src/include/optimizer/pathn=
ode.h
index a12af54971..37a946f857 100644
--- a/src/include/optimizer/pathnode.h
+++ b/src/include/optimizer/pathnode.h
@@ -44,6 +44,7 @@ extern IndexPath *create_index_path(PlannerInfo *root,
=09=09=09=09=09=09=09=09=09List *indexorderbys,
=09=09=09=09=09=09=09=09=09List *indexorderbycols,
=09=09=09=09=09=09=09=09=09List *pathkeys,
+=09=09=09=09=09=09=09=09=09List *uniquekeys,
=09=09=09=09=09=09=09=09=09ScanDirection indexscandir,
=09=09=09=09=09=09=09=09=09bool indexonly,
=09=09=09=09=09=09=09=09=09Relids required_outer,
diff --git a/src/include/optimizer/paths.h b/src/include/optimizer/paths.h
index c6c34630c2..c79e47eeaf 100644
--- a/src/include/optimizer/paths.h
+++ b/src/include/optimizer/paths.h
@@ -214,6 +214,9 @@ extern List *build_join_pathkeys(PlannerInfo *root,
extern List *make_pathkeys_for_sortclauses(PlannerInfo *root,
=09=09=09=09=09=09=09=09=09=09 List *sortclauses,
=09=09=09=09=09=09=09=09=09=09 List *tlist);
+extern List *make_pathkeys_for_uniquekeys(PlannerInfo *root,
+=09=09=09=09=09=09=09=09=09=09 List *sortclauses,
+=09=09=09=09=09=09=09=09=09=09 List *tlist);
extern void initialize_mergeclause_eclasses(PlannerInfo *root,
=09=09=09=09=09=09=09=09=09=09=09RestrictInfo *restrictinfo);
extern void update_mergeclause_eclasses(PlannerInfo *root,
@@ -240,4 +243,12 @@ extern PathKey *make_canonical_pathkey(PlannerInfo *ro=
ot,
extern void add_paths_to_append_rel(PlannerInfo *root, RelOptInfo *rel,
=09=09=09=09=09=09=09=09=09List *live_childrels);
=20
+/*
+ * uniquekey.c
+ *=09 Utilities for matching and building unique keys
+ */
+extern List *build_uniquekeys(PlannerInfo *root, List *sortclauses);
+extern bool uniquekeys_contained_in(List *keys1, List *keys2);
+extern bool has_useful_uniquekeys(PlannerInfo *root);
+
#endif=09=09=09=09=09=09=09/* PATHS_H */
--=20
2.21.0
--------------25BA7482CCCC3CA4CEC0D330
Content-Type: text/x-patch; charset=UTF-8;
name="v29_0002-Index-skip-scan.patch"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment;
filename="v29_0002-Index-skip-scan.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-25 11:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-25 11:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 196 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 209 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..5bb11add3c 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,190 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+ if (wc->rpSkipTo != ST_NEXT_ROW)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("only AFTER MATCH SKIP TO NEXT_ROW is supported")));
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Jun_25_21_05_09_2023_126)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v1-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-25 11:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-25 11:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 196 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 209 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..5bb11add3c 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,190 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+ if (wc->rpSkipTo != ST_NEXT_ROW)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("only AFTER MATCH SKIP TO NEXT_ROW is supported")));
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Jun_25_21_05_09_2023_126)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v1-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-25 11:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-25 11:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 196 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 209 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..5bb11add3c 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,190 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+ if (wc->rpSkipTo != ST_NEXT_ROW)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("only AFTER MATCH SKIP TO NEXT_ROW is supported")));
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Jun_25_21_05_09_2023_126)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v1-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-25 11:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-25 11:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 196 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 209 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..5bb11add3c 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,190 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+ if (wc->rpSkipTo != ST_NEXT_ROW)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("only AFTER MATCH SKIP TO NEXT_ROW is supported")));
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Jun_25_21_05_09_2023_126)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v1-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-25 11:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-25 11:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 196 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 209 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..5bb11add3c 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,190 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+ if (wc->rpSkipTo != ST_NEXT_ROW)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("only AFTER MATCH SKIP TO NEXT_ROW is supported")));
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Jun_25_21_05_09_2023_126)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v1-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-26 08:05 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-26 08:05 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 192 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 205 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..ccf3332bef 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,186 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jun_26_17_45_07_2023_724)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v2-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-26 08:05 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-26 08:05 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 192 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 205 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..ccf3332bef 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,186 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jun_26_17_45_07_2023_724)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v2-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-26 08:05 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-26 08:05 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 192 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 205 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..ccf3332bef 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,186 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jun_26_17_45_07_2023_724)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v2-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-26 08:05 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-26 08:05 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 192 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 205 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..ccf3332bef 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,186 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jun_26_17_45_07_2023_724)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v2-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-06-26 08:05 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-06-26 08:05 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 192 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 205 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index f61f794755..ccf3332bef 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2949,6 +2952,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3814,3 +3821,186 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static List *
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ List *patterns;
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return NULL;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+ return patterns;
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 346fd272b6..20231d9ec0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -541,6 +541,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1754,6 +1755,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3133,6 +3135,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jun_26_17_45_07_2023_724)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v2-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v3 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-07-26 10:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-07-26 10:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 190 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 203 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..ea2decc579 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,184 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Jul_26_21_21_34_2023_317)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v3-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v3 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-07-26 10:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-07-26 10:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 190 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 203 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..ea2decc579 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,184 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Jul_26_21_21_34_2023_317)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v3-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v3 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-07-26 10:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-07-26 10:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 190 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 203 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..ea2decc579 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,184 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Jul_26_21_21_34_2023_317)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v3-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v3 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-07-26 10:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-07-26 10:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 ++
src/backend/parser/parse_clause.c | 190 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 203 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..ea2decc579 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,184 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ /* Check Frame option. Frame must start at current row */
+
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+ ResTarget *restarget, *r;
+ List *restargets;
+
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char *name;
+ ListCell *l;
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Make sure that row pattern definition search condition is a boolean
+ * expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Jul_26_21_21_34_2023_317)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v3-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-08-09 07:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-08-09 07:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Aug__9_17_41_12_2023_134)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v4-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-08-09 07:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-08-09 07:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Aug__9_17_41_12_2023_134)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v4-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-08-09 07:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-08-09 07:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Aug__9_17_41_12_2023_134)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v4-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-08-09 07:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-08-09 07:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Aug__9_17_41_12_2023_134)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v4-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-08-09 07:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-08-09 07:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Aug__9_17_41_12_2023_134)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v4-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-08-09 07:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-08-09 07:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index fed8e4d089..8921b7ae01 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Aug__9_17_41_12_2023_134)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v4-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v5-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-12 05:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-12 05:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 305 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Sep_12_15_18_43_2023_359)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v6-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-22 04:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-22 04:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Sep_22_14_16_40_2023_530)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v7-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-22 04:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-22 04:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Sep_22_14_16_40_2023_530)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v7-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-22 04:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-22 04:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Sep_22_14_16_40_2023_530)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v7-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-22 04:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-22 04:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Sep_22_14_16_40_2023_530)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v7-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-22 04:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-22 04:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Sep_22_14_16_40_2023_530)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v7-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v8 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-25 05:01 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-25 05:01 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Sep_25_14_26_30_2023_752)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v8-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v8 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-25 05:01 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-25 05:01 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Sep_25_14_26_30_2023_752)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v8-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v8 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-25 05:01 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-25 05:01 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Sep_25_14_26_30_2023_752)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v8-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v8 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-25 05:01 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-09-25 05:01 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Sep_25_14_26_30_2023_752)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v8-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v9 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-04 05:51 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-04 05:51 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Oct__4_15_03_28_2023_821)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v9-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v9 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-04 05:51 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-04 05:51 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Oct__4_15_03_28_2023_821)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v9-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v9 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-04 05:51 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-04 05:51 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Oct__4_15_03_28_2023_821)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v9-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v9 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-04 05:51 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-04 05:51 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 293 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 306 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..293d4b1680 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Oct__4_15_03_28_2023_821)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v9-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v10 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-22 02:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-22 02:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 299 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..9c347216f7 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,293 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Oct_22_11_39_20_2023_140)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v10-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v10 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-22 02:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-22 02:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 299 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..9c347216f7 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,293 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Oct_22_11_39_20_2023_140)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v10-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v10 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-22 02:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-22 02:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 299 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..9c347216f7 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,293 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Oct_22_11_39_20_2023_140)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v10-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v10 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-10-22 02:22 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-10-22 02:22 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 299 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..9c347216f7 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,293 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc, *l;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /*
+ * Primary row pattern variable names in PATTERN clause must appear in
+ * DEFINE clause as row pattern definition variable names.
+ */
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ ResTarget *restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)a))));
+ }
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Oct_22_11_39_20_2023_140)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v10-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v11 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-11-08 06:57 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-11-08 06:57 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 278 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 291 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9bbad33fbd..28fb5e0d71 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..c5d3c10683 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,272 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 6c29471bb3..086431f91b 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Nov__8_16_37_05_2023_872)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v11-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v11 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-11-08 06:57 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-11-08 06:57 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 278 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 291 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9bbad33fbd..28fb5e0d71 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..c5d3c10683 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,272 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 6c29471bb3..086431f91b 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Nov__8_16_37_05_2023_872)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v11-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v11 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-11-08 06:57 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-11-08 06:57 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 278 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 291 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9bbad33fbd..28fb5e0d71 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..c5d3c10683 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,272 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 6c29471bb3..086431f91b 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_Nov__8_16_37_05_2023_872)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v11-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v12 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-12-04 11:23 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-12-04 11:23 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 278 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 291 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9bbad33fbd..28fb5e0d71 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..c5d3c10683 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,272 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 6c29471bb3..086431f91b 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Dec__8_10_16_13_2023_489)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v12-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v12 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-12-04 11:23 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-12-04 11:23 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 278 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 291 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9bbad33fbd..28fb5e0d71 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..c5d3c10683 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,272 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 6c29471bb3..086431f91b 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Dec__8_10_16_13_2023_489)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v12-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v12 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-12-04 11:23 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2023-12-04 11:23 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 278 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 291 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9bbad33fbd..28fb5e0d71 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..c5d3c10683 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,272 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 6c29471bb3..086431f91b 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Dec__8_10_16_13_2023_489)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v12-0003-Row-pattern-recognition-patch-planner.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v13 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-01-22 09:45 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-01-22 09:45 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 281 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 294 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 7b211a7743..5a9743be6e 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4b50278fd0..104c0105c5 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,275 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 9300c7b9ab..da4f42677b 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index fdb3e6df33..bf07085a15 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jan_22_19_26_18_2024_011)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v13-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v13 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-01-22 09:45 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-01-22 09:45 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 281 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 294 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 7b211a7743..5a9743be6e 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4b50278fd0..104c0105c5 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,275 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 9300c7b9ab..da4f42677b 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index fdb3e6df33..bf07085a15 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jan_22_19_26_18_2024_011)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v13-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v13 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-01-22 09:45 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-01-22 09:45 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 281 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 294 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 7b211a7743..5a9743be6e 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -575,6 +575,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -964,6 +968,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4b50278fd0..104c0105c5 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3822,275 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc, *l;
+ ResTarget *restarget, *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist.
+ * (the raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *)n;
+ restarget->location = -1;
+ restargets = lappend((List *)restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+ restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *)r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial.
+ * We assign [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *)lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+ parser_errposition(pstate, exprLocation((Node *)restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *)defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *)lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s","MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 9300c7b9ab..da4f42677b 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index fdb3e6df33..bf07085a15 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Jan_22_19_26_18_2024_011)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v13-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v14 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-02-28 13:59 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-02-28 13:59 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9d151a880b..a36218b103 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -576,6 +576,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -965,6 +969,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4b50278fd0..8a4f8f24d2 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2957,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3826,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 9300c7b9ab..da4f42677b 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index fdb3e6df33..bf07085a15 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Feb_29_09_19_54_2024_640)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v14-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v14 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-02-28 13:59 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-02-28 13:59 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9d151a880b..a36218b103 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -576,6 +576,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -965,6 +969,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4b50278fd0..8a4f8f24d2 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2957,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3826,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 9300c7b9ab..da4f42677b 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index fdb3e6df33..bf07085a15 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Feb_29_09_19_54_2024_640)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v14-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v14 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-02-28 13:59 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-02-28 13:59 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 9d151a880b..a36218b103 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -576,6 +576,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -965,6 +969,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4b50278fd0..8a4f8f24d2 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2950,6 +2957,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3815,3 +3826,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 9300c7b9ab..da4f42677b 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index fdb3e6df33..bf07085a15 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Feb_29_09_19_54_2024_640)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v14-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v15 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-03-28 10:30 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-03-28 10:30 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index d2ac86777c..de2e5791e3 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2948,6 +2955,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3813,3 +3824,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 73c83cea4a..8b0cc608bc 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Mar_28_19_59_25_2024_076)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v15-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v15 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-03-28 10:30 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-03-28 10:30 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index d2ac86777c..de2e5791e3 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2948,6 +2955,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3813,3 +3824,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 73c83cea4a..8b0cc608bc 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Mar_28_19_59_25_2024_076)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v15-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v15 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-03-28 10:30 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-03-28 10:30 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index d2ac86777c..de2e5791e3 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2948,6 +2955,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3813,3 +3824,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 73c83cea4a..8b0cc608bc 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Mar_28_19_59_25_2024_076)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v15-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v15 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-03-28 10:30 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-03-28 10:30 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index d2ac86777c..de2e5791e3 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2948,6 +2955,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3813,3 +3824,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 73c83cea4a..8b0cc608bc 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Mar_28_19_59_25_2024_076)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v15-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v15 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-03-28 10:30 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-03-28 10:30 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index d2ac86777c..de2e5791e3 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2948,6 +2955,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3813,3 +3824,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 73c83cea4a..8b0cc608bc 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Mar_28_19_59_25_2024_076)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v15-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v16 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-04-12 06:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-04-12 06:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4fc5fc87e0..003a1e14ce 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3821,3 +3832,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 4c98d7a046..a9f9d47854 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Apr_12_16_09_08_2024_262)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v16-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v16 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-04-12 06:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-04-12 06:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4fc5fc87e0..003a1e14ce 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3821,3 +3832,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 4c98d7a046..a9f9d47854 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Apr_12_16_09_08_2024_262)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v16-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v16 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-04-12 06:49 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-04-12 06:49 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4fc5fc87e0..003a1e14ce 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3821,3 +3832,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 4c98d7a046..a9f9d47854 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Apr_12_16_09_08_2024_262)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v16-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v17 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-04-28 11:00 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-04-28 11:00 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4fc5fc87e0..003a1e14ce 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3821,3 +3832,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 1c1c86aa3e..5540a0fb0a 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Apr_28_20_28_26_2024_444)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v17-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v17 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-04-28 11:00 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-04-28 11:00 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4fc5fc87e0..003a1e14ce 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3821,3 +3832,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 1c1c86aa3e..5540a0fb0a 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Apr_28_20_28_26_2024_444)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v17-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v17 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-04-28 11:00 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-04-28 11:00 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4fc5fc87e0..003a1e14ce 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->runCondition = NIL;
wc->winref = winref;
@@ -3821,3 +3832,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 1c1c86aa3e..5540a0fb0a 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 0cbc950c95..ad982a7c17 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2657,6 +2657,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sun_Apr_28_20_28_26_2024_444)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v17-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v18 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-11 07:11 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-11 07:11 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..eb138087bf 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_May_11_16_23_07_2024_789)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v18-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v18 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-11 07:11 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-11 07:11 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..eb138087bf 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_May_11_16_23_07_2024_789)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v18-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v18 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-11 07:11 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-11 07:11 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 4 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 309 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..eb138087bf 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1817,6 +1818,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_VALUES:
case EXPR_KIND_VALUES_SINGLE:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3197,6 +3199,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_May_11_16_23_07_2024_789)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v18-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v19 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-14 23:26 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-14 23:26 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..e98b45e06e 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_May_15_09_02_03_2024_008)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v19-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v19 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-14 23:26 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-14 23:26 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..e98b45e06e 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_May_15_09_02_03_2024_008)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v19-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v19 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-14 23:26 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-14 23:26 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..e98b45e06e 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Wed_May_15_09_02_03_2024_008)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v19-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v20 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-24 02:26 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-24 02:26 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..e98b45e06e 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_May_24_11_39_19_2024_763)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v20-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v20 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-24 02:26 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-24 02:26 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..e98b45e06e 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_May_24_11_39_19_2024_763)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v20-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v20 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-05-24 02:26 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-05-24 02:26 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index aba3546ed1..e98b45e06e 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -578,6 +578,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_May_24_11_39_19_2024_763)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v20-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v21 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-08-26 04:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-08-26 04:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 56e413da9f..c187b3278d 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Aug_26_13_39_47_2024_878)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v21-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v21 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-08-26 04:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-08-26 04:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 56e413da9f..c187b3278d 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Aug_26_13_39_47_2024_878)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v21-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v21 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-08-26 04:32 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-08-26 04:32 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index bee7d8346a..9bc22a836a 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -577,6 +577,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -967,6 +971,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 56e413da9f..c187b3278d 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Aug_26_13_39_47_2024_878)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v21-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v22 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-09-19 04:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-09-19 04:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index efa730c167..a80263f90d 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 36c1b7a88f..fe154bcaa0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Sep_19_13_59_47_2024_608)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v22-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v22 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-09-19 04:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-09-19 04:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index efa730c167..a80263f90d 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 36c1b7a88f..fe154bcaa0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Sep_19_13_59_47_2024_608)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v22-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v22 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-09-19 04:48 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-09-19 04:48 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 296 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 311 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index efa730c167..a80263f90d 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 8118036495..9762dce81f 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3820,3 +3831,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ * list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 36c1b7a88f..fe154bcaa0 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Sep_19_13_59_47_2024_608)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v22-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-02 14:46 Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-02 14:46 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>; Amit Langote <[email protected]>
On Wed, Oct 2, 2024 at 9:51 AM jian he <[email protected]> wrote:
>
> > >
> > > + typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
> > > + if (typentry == NULL || typentry->type_id != elmtyp)
> > > + {
> > > + typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > TYPECACHE_GT_OPR);
> > > + fcinfo->flinfo->fn_extra = (void *) typentry;
> > > + }
> > > you need to one-time check typentry->lt_opr or typentry->gt_opr exists?
> > > see CreateStatistics.
> > > /* Disallow data types without a less-than operator */
> > > type = lookup_type_cache(attForm->atttypid, TYPECACHE_LT_OPR);
> > > if (type->lt_opr == InvalidOid)
> > > ereport(ERROR,
> > > (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
> > > errmsg("column \"%s\" cannot be used in
> > > statistics because its type %s has no default btree operator class",
> > > attname, format_type_be(attForm->atttypid))));
> >
> > I added an Assert for this part, not sure if that is enough.
> >
>
> i think it really should be:
>
> if (typentry == NULL || typentry->type_id != elmtyp)
> {
> typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> TYPECACHE_GT_OPR);
> fcinfo->flinfo->fn_extra = (void *) typentry;
> if ((sort_asc && !OidIsValid(typentry->lt_opr) || (!sort_as &&
> OidIsValid(typentry->gt_opr));
> ereport(ERROR,....)
> }
>
> Imagine a type that doesn't have TYPECACHE_LT_OPR or TYPECACHE_GT_OPR
> then we cannot do the sort, we should just error out.
>
> I just tried this colour type [1] with (CREATE TYPE colour (INPUT =
> colour_in, OUTPUT = colour_out, LIKE = pg_catalog.int4);
>
> select array_sort('{#FF0000, #FF0000}'::colour[]);
> of course it will segfault with your new Assert.
>
>
> [1] https://github.com/hlinnaka/colour-datatype/blob/master/colour.c
Make sense, PFA v5 with Jian's suggestion.
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v5-0001-general-purpose-array_sort.patch (19.9K, ../../CAEG8a3KUiAM6_DSQHQb6Ox0Og1S7-sT5h45WJsmkF_Y-tGHKoQ@mail.gmail.com/2-v5-0001-general-purpose-array_sort.patch)
download | inline diff:
From e2ce82f1bc8326dd1a451b17c6df4cd61907c558 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v5] general purpose array_sort
Sorts anyarray in either ascending or descending order.
The array must be empty or one-dimensional.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 32 +++
src/backend/utils/adt/array_userfuncs.c | 186 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 6 +
src/test/regress/expected/arrays.out | 125 ++++++++++++
.../regress/expected/collate.icu.utf8.out | 13 ++
src/test/regress/sql/arrays.sql | 25 +++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 392 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index e39d524b6b..cf30484e2a 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20419,6 +20419,38 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional> <optional> , <parameter>dir</parameter> </optional>)
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the array based on the given parameter. The array must be empty or one-dimensional.
+ </para>
+ <para>
+ If the <literal>COLLATE</literal> option is specified then sorting is based on <replaceable>collation_name</replaceable>, otherwise
+ using array element type's collation.
+ <parameter>dir</parameter> must be one of the following, and that the starred options are the defaults when null position is omitted.
+<programlisting>
+ asc
+ desc
+ asc nulls first
+ asc nulls last *
+ desc nulls first *
+ desc nulls last
+ nulls first
+ nulls last
+</programlisting>
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[1,2,5,6,3,4])</literal>
+ <returnvalue>{1,2,3,4,5,6}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 6599be2ec5..9ec006d716 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -12,15 +12,18 @@
*/
#include "postgres.h"
+#include "catalog/namespace.h"
#include "catalog/pg_type.h"
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1685,3 +1688,186 @@ array_sample(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+
+#define WHITESPACE " \t\n\r\v\f"
+
+typedef enum
+{
+ PARSE_SORT_ORDER_INIT,
+ PARSE_SORT_ORDER_DIRECTION_SET,
+ PARSE_SORT_ORDER_NULLS_OPTION,
+ PARSE_SORT_ORDER_ERROR,
+ PARSE_SORT_ORDER_DONE
+} ParseSortOrderState;
+
+static bool
+parse_sort_order(const char *str, bool *sort_asc, bool *nulls_first)
+{
+ char *token;
+ char *saveptr;
+ char *str_copy = pstrdup(str);
+ bool nulls_first_set = false;
+ ParseSortOrderState state = PARSE_SORT_ORDER_INIT;
+
+ token = strtok_r(str_copy, WHITESPACE, &saveptr);
+
+ while (token != NULL && state != PARSE_SORT_ORDER_ERROR)
+ {
+ switch (state)
+ {
+ case PARSE_SORT_ORDER_INIT:
+ if (pg_strcasecmp(token, "ASC") == 0)
+ {
+ *sort_asc = true;
+ state = PARSE_SORT_ORDER_DIRECTION_SET;
+ }
+ else if (pg_strcasecmp(token, "DESC") == 0)
+ {
+ *sort_asc = false;
+ state = PARSE_SORT_ORDER_DIRECTION_SET;
+ }
+ else if (pg_strcasecmp(token, "NULLS") == 0)
+ state = PARSE_SORT_ORDER_NULLS_OPTION;
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_DIRECTION_SET:
+ if (pg_strcasecmp(token, "NULLS") == 0)
+ state = PARSE_SORT_ORDER_NULLS_OPTION;
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_NULLS_OPTION:
+ if (pg_strcasecmp(token, "FIRST") == 0)
+ {
+ *nulls_first = true;
+ nulls_first_set = true;
+ state = PARSE_SORT_ORDER_DONE;
+ }
+ else if (pg_strcasecmp(token, "LAST") == 0)
+ {
+ *nulls_first = false;
+ nulls_first_set = true;
+ state = PARSE_SORT_ORDER_DONE;
+ }
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_DONE:
+ /* No more tokens should be processed after first/last */
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ default:
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+ }
+
+ token = strtok_r(NULL, WHITESPACE, &saveptr);
+ }
+
+ if (state == PARSE_SORT_ORDER_INIT ||
+ state == PARSE_SORT_ORDER_DIRECTION_SET)
+ state = PARSE_SORT_ORDER_DONE;
+
+ if (state == PARSE_SORT_ORDER_NULLS_OPTION)
+ state = PARSE_SORT_ORDER_ERROR;
+
+ if (!nulls_first_set && state == PARSE_SORT_ORDER_DONE)
+ *nulls_first = !*sort_asc;
+
+ pfree(str_copy);
+ return state == PARSE_SORT_ORDER_DONE;
+}
+
+/*
+ * array_sort
+ *
+ * Sorts the array in either ascending or descending order.
+ * The array must be empty or one-dimensional.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ text *dirstr = (fcinfo->nargs > 1) ? PG_GETARG_TEXT_PP(1) : NULL;
+ bool sort_asc = true;
+ bool nulls_first = false;
+ Oid elmtyp;
+ Oid collation = PG_GET_COLLATION();
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildState *astate = NULL;
+
+ if (ARR_NDIM(array) > 1)
+ ereport(ERROR,
+ (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+ errmsg("multidimensional arrays sorting are not supported")));
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ if (dirstr != NULL)
+ {
+ if (!parse_sort_order(text_to_cstring(dirstr), &sort_asc, &nulls_first))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("second parameter must be a valid sort direction")));
+ }
+
+ elmtyp = ARR_ELEMTYPE(array);
+ typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((sort_asc && !OidIsValid(typentry->lt_opr)) ||
+ (!sort_asc && !OidIsValid(typentry->gt_opr)))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify an ordering operator for type %s",
+ format_type_be(elmtyp))));
+ fcinfo->flinfo->fn_extra = (void *) typentry;
+ }
+
+ tuplesortstate = tuplesort_begin_datum(elmtyp,
+ sort_asc ? typentry->lt_opr : typentry->gt_opr,
+ collation,
+ nulls_first, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, 0, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResult(astate, value, isnull,
+ elmtyp, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResult(astate, CurrentMemoryContext));
+}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 43f608d7a0..1b2d64ac39 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1734,6 +1734,12 @@
{ oid => '6216', descr => 'take samples from array',
proname => 'array_sample', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray int4', prosrc => 'array_sample' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray text', prosrc => 'array_sort_order'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index a6d81fd5f9..9c71a65b5a 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2703,3 +2703,128 @@ SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
ERROR: sample size must be between 0 and 6
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
ERROR: sample size must be between 0 and 6
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[], 'desc');
+ array_sort
+---------------
+ {6,5,4,3,2,1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'asc');
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'desc');
+ array_sort
+---------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[], 'desc');
+ array_sort
+---------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[]);
+ array_sort
+------------------------------------------------------------------
+ {"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[], 'desc');
+ array_sort
+------------------------------------------------------------------
+ {"DŽxxDŽ džxxDž Džxxdž",ȺȺȺ,ⱥⱥⱥ,ⱥȺ,"ábc sßs ßss DÉF","abc DEF 123abc"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[] COLLATE "pg_c_utf8", 'asc');
+ array_sort
+------------------------------------------------------------------
+ {"abc DEF 123abc","ábc sßs ßss DÉF","DŽxxDŽ džxxDž Džxxdž",ȺȺȺ,ⱥȺ,ⱥⱥⱥ}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[] COLLATE "pg_c_utf8", 'desc');
+ array_sort
+------------------------------------------------------------------
+ {ⱥⱥⱥ,ⱥȺ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž","ábc sßs ßss DÉF","abc DEF 123abc"}
+(1 row)
+
+-- nulls first/last tests
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[]);
+ array_sort
+----------------------------------------------------------------------------
+ {"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž",NULL,NULL}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'asc');
+ array_sort
+----------------------------------------------------------------------------
+ {"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž",NULL,NULL}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'desc');
+ array_sort
+----------------------------------------------------------------------------
+ {NULL,NULL,"DŽxxDŽ džxxDž Džxxdž",ȺȺȺ,ⱥⱥⱥ,ⱥȺ,"ábc sßs ßss DÉF","abc DEF 123abc"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'nulls first');
+ array_sort
+----------------------------------------------------------------------------
+ {NULL,NULL,"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'nulls first');
+ array_sort
+----------------------------------------------------------------------------
+ {NULL,NULL,"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'asc nulls first');
+ array_sort
+----------------------------------------------------------------------------
+ {NULL,NULL,"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'asc nulls last');
+ array_sort
+----------------------------------------------------------------------------
+ {"abc DEF 123abc","ábc sßs ßss DÉF",ⱥȺ,ⱥⱥⱥ,ȺȺȺ,"DŽxxDŽ džxxDž Džxxdž",NULL,NULL}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'desc nulls first');
+ array_sort
+----------------------------------------------------------------------------
+ {NULL,NULL,"DŽxxDŽ džxxDž Džxxdž",ȺȺȺ,ⱥⱥⱥ,ⱥȺ,"ábc sßs ßss DÉF","abc DEF 123abc"}
+(1 row)
+
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'desc nulls last');
+ array_sort
+----------------------------------------------------------------------------
+ {"DŽxxDŽ džxxDž Džxxdž",ȺȺȺ,ⱥⱥⱥ,ⱥȺ,"ábc sßs ßss DÉF","abc DEF 123abc",NULL,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[1,2],[3,4],[5,6]]); -- fail
+ERROR: multidimensional arrays sorting are not supported
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index 31345295c1..743b4d8199 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 47058dfde5..30606700ef 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -827,3 +827,28 @@ SELECT array_dims(array_sample('[-1:2][2:3]={{1,2},{3,NULL},{5,6},{7,8}}'::int[]
SELECT array_dims(array_sample('{{{1,2},{3,NULL}},{{5,6},{7,8}},{{9,10},{11,12}}}'::int[], 2));
SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[], 'desc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'asc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'desc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[], 'desc');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[]);
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[], 'desc');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[] COLLATE "pg_c_utf8", 'asc');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ}'::text[] COLLATE "pg_c_utf8", 'desc');
+-- nulls first/last tests
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[]);
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'asc');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'desc');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'nulls first');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'nulls first');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'asc nulls first');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'asc nulls last');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'desc nulls first');
+SELECT array_sort('{abc DEF 123abc,ábc sßs ßss DÉF,null,DŽxxDŽ džxxDž Džxxdž,ȺȺȺ,ⱥⱥⱥ,ⱥȺ,null}'::text[], 'desc nulls last');
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[1,2],[3,4],[5,6]]); -- fail
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index b6135f0347..99f1abe869 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2025,6 +2025,7 @@ ParseLoc
ParseNamespaceColumn
ParseNamespaceItem
ParseParamRefHook
+ParseSortOrderState
ParseState
ParsedLex
ParsedScript
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-03 06:22 Amit Langote <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Amit Langote @ 2024-10-03 06:22 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: jian he <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Junwang,
On Wed, Oct 2, 2024 at 11:46 PM Junwang Zhao <[email protected]> wrote:
> On Wed, Oct 2, 2024 at 9:51 AM jian he <[email protected]> wrote:
> >
> > > >
> > > > + typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
> > > > + if (typentry == NULL || typentry->type_id != elmtyp)
> > > > + {
> > > > + typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > > TYPECACHE_GT_OPR);
> > > > + fcinfo->flinfo->fn_extra = (void *) typentry;
> > > > + }
> > > > you need to one-time check typentry->lt_opr or typentry->gt_opr exists?
> > > > see CreateStatistics.
> > > > /* Disallow data types without a less-than operator */
> > > > type = lookup_type_cache(attForm->atttypid, TYPECACHE_LT_OPR);
> > > > if (type->lt_opr == InvalidOid)
> > > > ereport(ERROR,
> > > > (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
> > > > errmsg("column \"%s\" cannot be used in
> > > > statistics because its type %s has no default btree operator class",
> > > > attname, format_type_be(attForm->atttypid))));
> > >
> > > I added an Assert for this part, not sure if that is enough.
> > >
> >
> > i think it really should be:
> >
> > if (typentry == NULL || typentry->type_id != elmtyp)
> > {
> > typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > TYPECACHE_GT_OPR);
> > fcinfo->flinfo->fn_extra = (void *) typentry;
> > if ((sort_asc && !OidIsValid(typentry->lt_opr) || (!sort_as &&
> > OidIsValid(typentry->gt_opr));
> > ereport(ERROR,....)
> > }
> >
> > Imagine a type that doesn't have TYPECACHE_LT_OPR or TYPECACHE_GT_OPR
> > then we cannot do the sort, we should just error out.
> >
> > I just tried this colour type [1] with (CREATE TYPE colour (INPUT =
> > colour_in, OUTPUT = colour_out, LIKE = pg_catalog.int4);
> >
> > select array_sort('{#FF0000, #FF0000}'::colour[]);
> > of course it will segfault with your new Assert.
> >
> >
> > [1] https://github.com/hlinnaka/colour-datatype/blob/master/colour.c
>
> Make sense, PFA v5 with Jian's suggestion.
Have you noticed that the tests have failed on Cirrus CI runs of this patch?
https://cirrus-ci.com/github/postgresql-cfbot/postgresql/cf%2F5277
It might be related to the test machines having a different *default*
locale than your local environment, which could result in a different
sort order for the test data. You may need to add an explicit COLLATE
clause to the tests to ensure consistent sorting across systems.
--
Thanks, Amit Langote
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-09 14:10 Junwang Zhao <[email protected]>
parent: Amit Langote <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-09 14:10 UTC (permalink / raw)
To: Amit Langote <[email protected]>; +Cc: jian he <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Amit,
On Thu, Oct 3, 2024 at 2:22 PM Amit Langote <[email protected]> wrote:
>
> Hi Junwang,
>
> On Wed, Oct 2, 2024 at 11:46 PM Junwang Zhao <[email protected]> wrote:
> > On Wed, Oct 2, 2024 at 9:51 AM jian he <[email protected]> wrote:
> > >
> > > > >
> > > > > + typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
> > > > > + if (typentry == NULL || typentry->type_id != elmtyp)
> > > > > + {
> > > > > + typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > > > TYPECACHE_GT_OPR);
> > > > > + fcinfo->flinfo->fn_extra = (void *) typentry;
> > > > > + }
> > > > > you need to one-time check typentry->lt_opr or typentry->gt_opr exists?
> > > > > see CreateStatistics.
> > > > > /* Disallow data types without a less-than operator */
> > > > > type = lookup_type_cache(attForm->atttypid, TYPECACHE_LT_OPR);
> > > > > if (type->lt_opr == InvalidOid)
> > > > > ereport(ERROR,
> > > > > (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
> > > > > errmsg("column \"%s\" cannot be used in
> > > > > statistics because its type %s has no default btree operator class",
> > > > > attname, format_type_be(attForm->atttypid))));
> > > >
> > > > I added an Assert for this part, not sure if that is enough.
> > > >
> > >
> > > i think it really should be:
> > >
> > > if (typentry == NULL || typentry->type_id != elmtyp)
> > > {
> > > typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > TYPECACHE_GT_OPR);
> > > fcinfo->flinfo->fn_extra = (void *) typentry;
> > > if ((sort_asc && !OidIsValid(typentry->lt_opr) || (!sort_as &&
> > > OidIsValid(typentry->gt_opr));
> > > ereport(ERROR,....)
> > > }
> > >
> > > Imagine a type that doesn't have TYPECACHE_LT_OPR or TYPECACHE_GT_OPR
> > > then we cannot do the sort, we should just error out.
> > >
> > > I just tried this colour type [1] with (CREATE TYPE colour (INPUT =
> > > colour_in, OUTPUT = colour_out, LIKE = pg_catalog.int4);
> > >
> > > select array_sort('{#FF0000, #FF0000}'::colour[]);
> > > of course it will segfault with your new Assert.
> > >
> > >
> > > [1] https://github.com/hlinnaka/colour-datatype/blob/master/colour.c
> >
> > Make sense, PFA v5 with Jian's suggestion.
>
> Have you noticed that the tests have failed on Cirrus CI runs of this patch?
>
> https://cirrus-ci.com/github/postgresql-cfbot/postgresql/cf%2F5277
Sorry for the late reply due to my vacation. I should have paid
more attention to Cirrus CI earlier ;)
>
> It might be related to the test machines having a different *default*
> locale than your local environment, which could result in a different
> sort order for the test data. You may need to add an explicit COLLATE
> clause to the tests to ensure consistent sorting across systems.
I've changed the tests to use just ASCII characters, then added
*COLLATE "C"* to the tests and CI passed, PFA v6.
>
> --
> Thanks, Amit Langote
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v6-0001-general-purpose-array_sort.patch (16.3K, ../../CAEG8a3LyWk8EuQZTA=1h9EwNbysXZ0Sb=uYm+5Rc10O7o-aGVw@mail.gmail.com/2-v6-0001-general-purpose-array_sort.patch)
download | inline diff:
From e44100b2947305cba33dec6baa6d03398daf363b Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v6] general purpose array_sort
Sorts anyarray in either ascending or descending order.
The array must be empty or one-dimensional.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 32 +++
src/backend/utils/adt/array_userfuncs.c | 186 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 6 +
src/test/regress/expected/arrays.out | 119 +++++++++++
.../regress/expected/collate.icu.utf8.out | 13 ++
src/test/regress/sql/arrays.sql | 24 +++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 385 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 7b4fbb5047..49d3197955 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20421,6 +20421,38 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional> <optional> , <parameter>dir</parameter> </optional>)
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the array based on the given parameter. The array must be empty or one-dimensional.
+ </para>
+ <para>
+ If the <literal>COLLATE</literal> option is specified then sorting is based on <replaceable>collation_name</replaceable>, otherwise
+ using array element type's collation.
+ <parameter>dir</parameter> must be one of the following, and that the starred options are the defaults when null position is omitted.
+<programlisting>
+ asc
+ desc
+ asc nulls first
+ asc nulls last *
+ desc nulls first *
+ desc nulls last
+ nulls first
+ nulls last
+</programlisting>
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[1,2,5,6,3,4])</literal>
+ <returnvalue>{1,2,3,4,5,6}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 6599be2ec5..9ec006d716 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -12,15 +12,18 @@
*/
#include "postgres.h"
+#include "catalog/namespace.h"
#include "catalog/pg_type.h"
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1685,3 +1688,186 @@ array_sample(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+
+#define WHITESPACE " \t\n\r\v\f"
+
+typedef enum
+{
+ PARSE_SORT_ORDER_INIT,
+ PARSE_SORT_ORDER_DIRECTION_SET,
+ PARSE_SORT_ORDER_NULLS_OPTION,
+ PARSE_SORT_ORDER_ERROR,
+ PARSE_SORT_ORDER_DONE
+} ParseSortOrderState;
+
+static bool
+parse_sort_order(const char *str, bool *sort_asc, bool *nulls_first)
+{
+ char *token;
+ char *saveptr;
+ char *str_copy = pstrdup(str);
+ bool nulls_first_set = false;
+ ParseSortOrderState state = PARSE_SORT_ORDER_INIT;
+
+ token = strtok_r(str_copy, WHITESPACE, &saveptr);
+
+ while (token != NULL && state != PARSE_SORT_ORDER_ERROR)
+ {
+ switch (state)
+ {
+ case PARSE_SORT_ORDER_INIT:
+ if (pg_strcasecmp(token, "ASC") == 0)
+ {
+ *sort_asc = true;
+ state = PARSE_SORT_ORDER_DIRECTION_SET;
+ }
+ else if (pg_strcasecmp(token, "DESC") == 0)
+ {
+ *sort_asc = false;
+ state = PARSE_SORT_ORDER_DIRECTION_SET;
+ }
+ else if (pg_strcasecmp(token, "NULLS") == 0)
+ state = PARSE_SORT_ORDER_NULLS_OPTION;
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_DIRECTION_SET:
+ if (pg_strcasecmp(token, "NULLS") == 0)
+ state = PARSE_SORT_ORDER_NULLS_OPTION;
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_NULLS_OPTION:
+ if (pg_strcasecmp(token, "FIRST") == 0)
+ {
+ *nulls_first = true;
+ nulls_first_set = true;
+ state = PARSE_SORT_ORDER_DONE;
+ }
+ else if (pg_strcasecmp(token, "LAST") == 0)
+ {
+ *nulls_first = false;
+ nulls_first_set = true;
+ state = PARSE_SORT_ORDER_DONE;
+ }
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_DONE:
+ /* No more tokens should be processed after first/last */
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ default:
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+ }
+
+ token = strtok_r(NULL, WHITESPACE, &saveptr);
+ }
+
+ if (state == PARSE_SORT_ORDER_INIT ||
+ state == PARSE_SORT_ORDER_DIRECTION_SET)
+ state = PARSE_SORT_ORDER_DONE;
+
+ if (state == PARSE_SORT_ORDER_NULLS_OPTION)
+ state = PARSE_SORT_ORDER_ERROR;
+
+ if (!nulls_first_set && state == PARSE_SORT_ORDER_DONE)
+ *nulls_first = !*sort_asc;
+
+ pfree(str_copy);
+ return state == PARSE_SORT_ORDER_DONE;
+}
+
+/*
+ * array_sort
+ *
+ * Sorts the array in either ascending or descending order.
+ * The array must be empty or one-dimensional.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ text *dirstr = (fcinfo->nargs > 1) ? PG_GETARG_TEXT_PP(1) : NULL;
+ bool sort_asc = true;
+ bool nulls_first = false;
+ Oid elmtyp;
+ Oid collation = PG_GET_COLLATION();
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildState *astate = NULL;
+
+ if (ARR_NDIM(array) > 1)
+ ereport(ERROR,
+ (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+ errmsg("multidimensional arrays sorting are not supported")));
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ if (dirstr != NULL)
+ {
+ if (!parse_sort_order(text_to_cstring(dirstr), &sort_asc, &nulls_first))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("second parameter must be a valid sort direction")));
+ }
+
+ elmtyp = ARR_ELEMTYPE(array);
+ typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((sort_asc && !OidIsValid(typentry->lt_opr)) ||
+ (!sort_asc && !OidIsValid(typentry->gt_opr)))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify an ordering operator for type %s",
+ format_type_be(elmtyp))));
+ fcinfo->flinfo->fn_extra = (void *) typentry;
+ }
+
+ tuplesortstate = tuplesort_begin_datum(elmtyp,
+ sort_asc ? typentry->lt_opr : typentry->gt_opr,
+ collation,
+ nulls_first, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, 0, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResult(astate, value, isnull,
+ elmtyp, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResult(astate, CurrentMemoryContext));
+}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 77f54a79e6..5850ecf550 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1734,6 +1734,12 @@
{ oid => '6216', descr => 'take samples from array',
proname => 'array_sample', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray int4', prosrc => 'array_sample' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray text', prosrc => 'array_sort_order'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index a6d81fd5f9..426a34487e 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2703,3 +2703,122 @@ SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
ERROR: sample size must be between 0 and 6
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
ERROR: sample size must be between 0 and 6
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[], 'desc');
+ array_sort
+---------------
+ {6,5,4,3,2,1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'asc');
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'desc');
+ array_sort
+---------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[], 'desc');
+ array_sort
+---------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'asc');
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'desc');
+ array_sort
+-----------------------
+ {foo,bbc,bar,CCC,Abc}
+(1 row)
+
+-- nulls first/last tests
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc');
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc');
+ array_sort
+----------------------------
+ {NULL,foo,bbc,bar,CCC,Abc}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+ array_sort
+----------------------------
+ {NULL,Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+ array_sort
+----------------------------
+ {NULL,Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls first');
+ array_sort
+----------------------------
+ {NULL,Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls last');
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls first');
+ array_sort
+----------------------------
+ {NULL,foo,bbc,bar,CCC,Abc}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls last');
+ array_sort
+----------------------------
+ {foo,bbc,bar,CCC,Abc,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[1,2],[3,4],[5,6]]); -- fail
+ERROR: multidimensional arrays sorting are not supported
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index 31345295c1..743b4d8199 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 47058dfde5..dbc0a4e23e 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -827,3 +827,27 @@ SELECT array_dims(array_sample('[-1:2][2:3]={{1,2},{3,NULL},{5,6},{7,8}}'::int[]
SELECT array_dims(array_sample('{{{1,2},{3,NULL}},{{5,6},{7,8}},{{9,10},{11,12}}}'::int[], 2));
SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[], 'desc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'asc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'desc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[], 'desc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'asc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'desc');
+-- nulls first/last tests
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls last');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls last');
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[1,2],[3,4],[5,6]]); -- fail
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 5fabb127d7..c4c8a11bd3 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2025,6 +2025,7 @@ ParseLoc
ParseNamespaceColumn
ParseNamespaceItem
ParseParamRefHook
+ParseSortOrderState
ParseState
ParsedLex
ParsedScript
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-09 15:46 Junwang Zhao <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-09 15:46 UTC (permalink / raw)
To: Amit Langote <[email protected]>; +Cc: jian he <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Wed, Oct 9, 2024 at 10:10 PM Junwang Zhao <[email protected]> wrote:
>
> Hi Amit,
>
> On Thu, Oct 3, 2024 at 2:22 PM Amit Langote <[email protected]> wrote:
> >
> > Hi Junwang,
> >
> > On Wed, Oct 2, 2024 at 11:46 PM Junwang Zhao <[email protected]> wrote:
> > > On Wed, Oct 2, 2024 at 9:51 AM jian he <[email protected]> wrote:
> > > >
> > > > > >
> > > > > > + typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
> > > > > > + if (typentry == NULL || typentry->type_id != elmtyp)
> > > > > > + {
> > > > > > + typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > > > > TYPECACHE_GT_OPR);
> > > > > > + fcinfo->flinfo->fn_extra = (void *) typentry;
> > > > > > + }
> > > > > > you need to one-time check typentry->lt_opr or typentry->gt_opr exists?
> > > > > > see CreateStatistics.
> > > > > > /* Disallow data types without a less-than operator */
> > > > > > type = lookup_type_cache(attForm->atttypid, TYPECACHE_LT_OPR);
> > > > > > if (type->lt_opr == InvalidOid)
> > > > > > ereport(ERROR,
> > > > > > (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
> > > > > > errmsg("column \"%s\" cannot be used in
> > > > > > statistics because its type %s has no default btree operator class",
> > > > > > attname, format_type_be(attForm->atttypid))));
> > > > >
> > > > > I added an Assert for this part, not sure if that is enough.
> > > > >
> > > >
> > > > i think it really should be:
> > > >
> > > > if (typentry == NULL || typentry->type_id != elmtyp)
> > > > {
> > > > typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > > TYPECACHE_GT_OPR);
> > > > fcinfo->flinfo->fn_extra = (void *) typentry;
> > > > if ((sort_asc && !OidIsValid(typentry->lt_opr) || (!sort_as &&
> > > > OidIsValid(typentry->gt_opr));
> > > > ereport(ERROR,....)
> > > > }
> > > >
> > > > Imagine a type that doesn't have TYPECACHE_LT_OPR or TYPECACHE_GT_OPR
> > > > then we cannot do the sort, we should just error out.
> > > >
> > > > I just tried this colour type [1] with (CREATE TYPE colour (INPUT =
> > > > colour_in, OUTPUT = colour_out, LIKE = pg_catalog.int4);
> > > >
> > > > select array_sort('{#FF0000, #FF0000}'::colour[]);
> > > > of course it will segfault with your new Assert.
> > > >
> > > >
> > > > [1] https://github.com/hlinnaka/colour-datatype/blob/master/colour.c
> > >
> > > Make sense, PFA v5 with Jian's suggestion.
> >
> > Have you noticed that the tests have failed on Cirrus CI runs of this patch?
> >
> > https://cirrus-ci.com/github/postgresql-cfbot/postgresql/cf%2F5277
>
> Sorry for the late reply due to my vacation. I should have paid
> more attention to Cirrus CI earlier ;)
>
> >
> > It might be related to the test machines having a different *default*
> > locale than your local environment, which could result in a different
> > sort order for the test data. You may need to add an explicit COLLATE
> > clause to the tests to ensure consistent sorting across systems.
>
> I've changed the tests to use just ASCII characters, then added
> *COLLATE "C"* to the tests and CI passed, PFA v6.
Sadly the CI only passed on my own github repo, it failed on
cfbot[1], will dig into the reason later because I can not open the cirrus
ci page right now ;(
[1] https://cirrus-ci.com/task/5815925960605696
>
> >
> > --
> > Thanks, Amit Langote
>
>
>
> --
> Regards
> Junwang Zhao
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-10 02:30 Junwang Zhao <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-10 02:30 UTC (permalink / raw)
To: Amit Langote <[email protected]>; +Cc: jian he <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Wed, Oct 9, 2024 at 11:46 PM Junwang Zhao <[email protected]> wrote:
>
> On Wed, Oct 9, 2024 at 10:10 PM Junwang Zhao <[email protected]> wrote:
> >
> > Hi Amit,
> >
> > On Thu, Oct 3, 2024 at 2:22 PM Amit Langote <[email protected]> wrote:
> > >
> > > Hi Junwang,
> > >
> > > On Wed, Oct 2, 2024 at 11:46 PM Junwang Zhao <[email protected]> wrote:
> > > > On Wed, Oct 2, 2024 at 9:51 AM jian he <[email protected]> wrote:
> > > > >
> > > > > > >
> > > > > > > + typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
> > > > > > > + if (typentry == NULL || typentry->type_id != elmtyp)
> > > > > > > + {
> > > > > > > + typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > > > > > TYPECACHE_GT_OPR);
> > > > > > > + fcinfo->flinfo->fn_extra = (void *) typentry;
> > > > > > > + }
> > > > > > > you need to one-time check typentry->lt_opr or typentry->gt_opr exists?
> > > > > > > see CreateStatistics.
> > > > > > > /* Disallow data types without a less-than operator */
> > > > > > > type = lookup_type_cache(attForm->atttypid, TYPECACHE_LT_OPR);
> > > > > > > if (type->lt_opr == InvalidOid)
> > > > > > > ereport(ERROR,
> > > > > > > (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
> > > > > > > errmsg("column \"%s\" cannot be used in
> > > > > > > statistics because its type %s has no default btree operator class",
> > > > > > > attname, format_type_be(attForm->atttypid))));
> > > > > >
> > > > > > I added an Assert for this part, not sure if that is enough.
> > > > > >
> > > > >
> > > > > i think it really should be:
> > > > >
> > > > > if (typentry == NULL || typentry->type_id != elmtyp)
> > > > > {
> > > > > typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR :
> > > > > TYPECACHE_GT_OPR);
> > > > > fcinfo->flinfo->fn_extra = (void *) typentry;
> > > > > if ((sort_asc && !OidIsValid(typentry->lt_opr) || (!sort_as &&
> > > > > OidIsValid(typentry->gt_opr));
> > > > > ereport(ERROR,....)
> > > > > }
> > > > >
> > > > > Imagine a type that doesn't have TYPECACHE_LT_OPR or TYPECACHE_GT_OPR
> > > > > then we cannot do the sort, we should just error out.
> > > > >
> > > > > I just tried this colour type [1] with (CREATE TYPE colour (INPUT =
> > > > > colour_in, OUTPUT = colour_out, LIKE = pg_catalog.int4);
> > > > >
> > > > > select array_sort('{#FF0000, #FF0000}'::colour[]);
> > > > > of course it will segfault with your new Assert.
> > > > >
> > > > >
> > > > > [1] https://github.com/hlinnaka/colour-datatype/blob/master/colour.c
> > > >
> > > > Make sense, PFA v5 with Jian's suggestion.
> > >
> > > Have you noticed that the tests have failed on Cirrus CI runs of this patch?
> > >
> > > https://cirrus-ci.com/github/postgresql-cfbot/postgresql/cf%2F5277
> >
> > Sorry for the late reply due to my vacation. I should have paid
> > more attention to Cirrus CI earlier ;)
> >
> > >
> > > It might be related to the test machines having a different *default*
> > > locale than your local environment, which could result in a different
> > > sort order for the test data. You may need to add an explicit COLLATE
> > > clause to the tests to ensure consistent sorting across systems.
> >
> > I've changed the tests to use just ASCII characters, then added
> > *COLLATE "C"* to the tests and CI passed, PFA v6.
>
> Sadly the CI only passed on my own github repo, it failed on
> cfbot[1], will dig into the reason later because I can not open the cirrus
> ci page right now ;(
>
> [1] https://cirrus-ci.com/task/5815925960605696
>
Seems the failure is not related to this patch, I guess the reason for
this is the stop phase doesn't
release the port properly?
2024-10-09 14:53:10.079 UTC [43052][checkpointer] LOG: checkpoint
complete: wrote 5617 buffers (34.3%), wrote 3 SLRU buffers; 0 WAL
file(s) added, 0 removed, 3 recycled; write=0.107 s, sync=0.001 s,
total=0.107 s; sync files=0, longest=0.000 s, average=0.000 s;
distance=45239 kB, estimate=45239 kB; lsn=0/414A138, redo
lsn=0/414A138
2024-10-09 14:53:10.084 UTC [43050][postmaster] LOG: database system
is shut down
2024-10-09 14:53:10.215 UTC [43270][postmaster] LOG: starting
PostgreSQL 18devel on x86_64-freebsd, compiled by clang-17.0.6, 64-bit
2024-10-09 14:53:10.215 UTC [43270][postmaster] LOG: could not bind
IPv4 address "127.0.0.1": Address already in use
2024-10-09 14:53:10.215 UTC [43270][postmaster] HINT: Is another
postmaster already running on port 11643? If not, wait a few seconds
and retry.
2024-10-09 14:53:10.215 UTC [43270][postmaster] WARNING: could not
create listen socket for "127.0.0.1"
2024-10-09 14:53:10.218 UTC [43270][postmaster] FATAL: could not
create any TCP/IP sockets
2024-10-09 14:53:10.218 UTC [43270][postmaster] LOG: database system
is shut down
https://api.cirrus-ci.com/v1/artifact/task/5815925960605696/testrun/build/testrun/ssl/001_ssltests/l...
> >
> > >
> > > --
> > > Thanks, Amit Langote
> >
> >
> >
> > --
> > Regards
> > Junwang Zhao
>
>
>
> --
> Regards
> Junwang Zhao
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-11 05:12 jian he <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: jian he @ 2024-10-11 05:12 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: Amit Langote <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
tricky case:
should we allow array element type to be composite/domain?
currently seems to work fine.
create table t(b int[]);
insert into t values ('{{1,3}}'), ('{{1,2}}');
select array_sort((select array_agg(t) from t), 'desc');
array_sort
-----------------------------------
{"(\"{{1,3}}\")","(\"{{1,2}}\")"}
select array_sort((t.b)) from t;
ERROR: multidimensional arrays sorting are not supported
select array_sort((select array_agg(t.b) from t));
ERROR: multidimensional arrays sorting are not supported
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-13 14:43 Junwang Zhao <[email protected]>
parent: jian he <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-13 14:43 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Amit Langote <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Jian,
On Fri, Oct 11, 2024 at 1:12 PM jian he <[email protected]> wrote:
>
> tricky case:
> should we allow array element type to be composite/domain?
> currently seems to work fine.
>
>
> create table t(b int[]);
> insert into t values ('{{1,3}}'), ('{{1,2}}');
>
> select array_sort((select array_agg(t) from t), 'desc');
> array_sort
> -----------------------------------
> {"(\"{{1,3}}\")","(\"{{1,2}}\")"}
>
>
> select array_sort((t.b)) from t;
> ERROR: multidimensional arrays sorting are not supported
>
>
> select array_sort((select array_agg(t.b) from t));
> ERROR: multidimensional arrays sorting are not supported
I tried the above cases, and the first one works because it's
a one dim array of composite type, the other two fails because
they are multidimensional.
It seems there is not much meaning to sort composite type,
so are you proposing we should error on that?
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-13 20:10 Tom Lane <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 2 replies; 166+ messages in thread
From: Tom Lane @ 2024-10-13 20:10 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: jian he <[email protected]>; Amit Langote <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Junwang Zhao <[email protected]> writes:
> It seems there is not much meaning to sort composite type,
> so are you proposing we should error on that?
It's hardly "general purpose" if it randomly refuses to
sort certain types. I would say it should be able to sort
anything that ORDER BY will handle --- and that certainly
includes the cases shown here.
regards, tom lane
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-23 14:28 Junwang Zhao <[email protected]>
parent: Tom Lane <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-23 14:28 UTC (permalink / raw)
To: Tom Lane <[email protected]>; +Cc: jian he <[email protected]>; Amit Langote <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Sorry for the late reply.
On Mon, Oct 14, 2024 at 4:10 AM Tom Lane <[email protected]> wrote:
>
> Junwang Zhao <[email protected]> writes:
> > It seems there is not much meaning to sort composite type,
> > so are you proposing we should error on that?
>
> It's hardly "general purpose" if it randomly refuses to
> sort certain types. I would say it should be able to sort
> anything that ORDER BY will handle --- and that certainly
> includes the cases shown here.
Yeah, agreed.
PFA v7 with multi-array support.
>
> regards, tom lane
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v7-0001-general-purpose-array_sort.patch (16.9K, ../../CAEG8a3L7RkH1f1Npru_VZ6_2hC1wJCboVWVpjKzfrY4P_=kvxg@mail.gmail.com/2-v7-0001-general-purpose-array_sort.patch)
download | inline diff:
From 783791442fdb4005cdfb4775f8aad977cbfbfce9 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v7] general purpose array_sort
Sorts anyarray in either ascending or descending order.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 32 +++
src/backend/utils/adt/array_userfuncs.c | 205 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 6 +
src/test/regress/expected/arrays.out | 123 +++++++++++
.../regress/expected/collate.icu.utf8.out | 13 ++
src/test/regress/sql/arrays.sql | 24 ++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 408 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index ad663c94d7..0414b43523 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20421,6 +20421,38 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional> <optional> , <parameter>dir</parameter> </optional>)
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the array based on the given parameter.
+ </para>
+ <para>
+ If the <literal>COLLATE</literal> option is specified then sorting is based on <replaceable>collation_name</replaceable>, otherwise
+ using array element type's collation.
+ <parameter>dir</parameter> must be one of the following, and that the starred options are the defaults when null position is omitted.
+<programlisting>
+ asc
+ desc
+ asc nulls first
+ asc nulls last *
+ desc nulls first *
+ desc nulls last
+ nulls first
+ nulls last
+</programlisting>
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 6599be2ec5..b6c948e60f 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -12,15 +12,18 @@
*/
#include "postgres.h"
+#include "catalog/namespace.h"
#include "catalog/pg_type.h"
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1685,3 +1688,205 @@ array_sample(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+
+#define WHITESPACE " \t\n\r\v\f"
+
+typedef enum
+{
+ PARSE_SORT_ORDER_INIT,
+ PARSE_SORT_ORDER_DIRECTION_SET,
+ PARSE_SORT_ORDER_NULLS_OPTION,
+ PARSE_SORT_ORDER_ERROR,
+ PARSE_SORT_ORDER_DONE
+} ParseSortOrderState;
+
+static bool
+parse_sort_order(const char *str, bool *sort_asc, bool *nulls_first)
+{
+ char *token;
+ char *saveptr;
+ char *str_copy = pstrdup(str);
+ bool nulls_first_set = false;
+ ParseSortOrderState state = PARSE_SORT_ORDER_INIT;
+
+ token = strtok_r(str_copy, WHITESPACE, &saveptr);
+
+ while (token != NULL && state != PARSE_SORT_ORDER_ERROR)
+ {
+ switch (state)
+ {
+ case PARSE_SORT_ORDER_INIT:
+ if (pg_strcasecmp(token, "ASC") == 0)
+ {
+ *sort_asc = true;
+ state = PARSE_SORT_ORDER_DIRECTION_SET;
+ }
+ else if (pg_strcasecmp(token, "DESC") == 0)
+ {
+ *sort_asc = false;
+ state = PARSE_SORT_ORDER_DIRECTION_SET;
+ }
+ else if (pg_strcasecmp(token, "NULLS") == 0)
+ state = PARSE_SORT_ORDER_NULLS_OPTION;
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_DIRECTION_SET:
+ if (pg_strcasecmp(token, "NULLS") == 0)
+ state = PARSE_SORT_ORDER_NULLS_OPTION;
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_NULLS_OPTION:
+ if (pg_strcasecmp(token, "FIRST") == 0)
+ {
+ *nulls_first = true;
+ nulls_first_set = true;
+ state = PARSE_SORT_ORDER_DONE;
+ }
+ else if (pg_strcasecmp(token, "LAST") == 0)
+ {
+ *nulls_first = false;
+ nulls_first_set = true;
+ state = PARSE_SORT_ORDER_DONE;
+ }
+ else
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ case PARSE_SORT_ORDER_DONE:
+ /* No more tokens should be processed after first/last */
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+
+ default:
+ state = PARSE_SORT_ORDER_ERROR;
+ break;
+ }
+
+ token = strtok_r(NULL, WHITESPACE, &saveptr);
+ }
+
+ if (state == PARSE_SORT_ORDER_INIT ||
+ state == PARSE_SORT_ORDER_DIRECTION_SET)
+ state = PARSE_SORT_ORDER_DONE;
+
+ if (state == PARSE_SORT_ORDER_NULLS_OPTION)
+ state = PARSE_SORT_ORDER_ERROR;
+
+ if (!nulls_first_set && state == PARSE_SORT_ORDER_DONE)
+ *nulls_first = !*sort_asc;
+
+ pfree(str_copy);
+ return state == PARSE_SORT_ORDER_DONE;
+}
+
+/*
+ * array_sort
+ *
+ * Sorts the array in either ascending or descending order.
+ * The array must be empty or one-dimensional.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ text *dirstr = (fcinfo->nargs > 1) ? PG_GETARG_TEXT_PP(1) : NULL;
+ bool sort_asc = true;
+ bool nulls_first = false;
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ if (dirstr != NULL)
+ {
+ if (!parse_sort_order(text_to_cstring(dirstr), &sort_asc, &nulls_first))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("second parameter must be a valid sort direction")));
+ }
+
+ elmtyp = ARR_ELEMTYPE(array);
+ typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
+ if (ARR_NDIM(array) == 1)
+ {
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, sort_asc ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((sort_asc && !OidIsValid(typentry->lt_opr)) ||
+ (!sort_asc && !OidIsValid(typentry->gt_opr)))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify an ordering operator for type %s",
+ format_type_be(elmtyp))));
+ fcinfo->flinfo->fn_extra = (void *) typentry;
+ }
+ }
+ else
+ {
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, sort_asc ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((sort_asc && !OidIsValid(typentry->lt_opr)) ||
+ (!sort_asc && !OidIsValid(typentry->gt_opr)))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify an ordering operator for type %s",
+ format_type_be(elmtyp))));
+ fcinfo->flinfo->fn_extra = (void *) typentry;
+ }
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ sort_asc ? typentry->lt_opr : typentry->gt_opr,
+ collation,
+ nulls_first, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 7c0b74fe05..a92fd3a6c2 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1741,6 +1741,12 @@
{ oid => '6216', descr => 'take samples from array',
proname => 'array_sample', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray int4', prosrc => 'array_sample' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray text', prosrc => 'array_sort_order'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index a6d81fd5f9..3c733e5aab 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2703,3 +2703,126 @@ SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
ERROR: sample size must be between 0 and 6
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
ERROR: sample size must be between 0 and 6
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[], 'desc');
+ array_sort
+---------------
+ {6,5,4,3,2,1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'asc');
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'desc');
+ array_sort
+---------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[], 'desc');
+ array_sort
+---------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'asc');
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'desc');
+ array_sort
+-----------------------
+ {foo,bbc,bar,CCC,Abc}
+(1 row)
+
+-- nulls first/last tests
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc');
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc');
+ array_sort
+----------------------------
+ {NULL,foo,bbc,bar,CCC,Abc}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+ array_sort
+----------------------------
+ {NULL,Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+ array_sort
+----------------------------
+ {NULL,Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls first');
+ array_sort
+----------------------------
+ {NULL,Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls last');
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls first');
+ array_sort
+----------------------------
+ {NULL,foo,bbc,bar,CCC,Abc}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls last');
+ array_sort
+----------------------------
+ {foo,bbc,bar,CCC,Abc,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 47058dfde5..821c4eab0b 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -827,3 +827,27 @@ SELECT array_dims(array_sample('[-1:2][2:3]={{1,2},{3,NULL},{5,6},{7,8}}'::int[]
SELECT array_dims(array_sample('{{{1,2},{3,NULL}},{{5,6},{7,8}},{{9,10},{11,12}}}'::int[], 2));
SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[], 'desc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'asc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::float8[], 'desc');
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[], 'desc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'asc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C", 'desc');
+-- nulls first/last tests
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'asc nulls last');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls first');
+SELECT array_sort('{foo,bar,CCC,Abc,bbc,null}'::text[] COLLATE "C", 'desc nulls last');
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 57de1acff3..f085b3cf87 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2025,6 +2025,7 @@ ParseLoc
ParseNamespaceColumn
ParseNamespaceItem
ParseParamRefHook
+ParseSortOrderState
ParseState
ParsedLex
ParsedScript
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 12:39 jian he <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: jian he @ 2024-10-24 12:39 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: Tom Lane <[email protected]>; Amit Langote <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Wed, Oct 23, 2024 at 10:28 PM Junwang Zhao <[email protected]> wrote:
> PFA v7 with multi-array support.
>
if (ARR_NDIM(array) == 1)
{
}
else
{
}
can be simplified.
i think beginning part of array_sort can be like the following:
(newline emitted)
---------------------------------------------------------------------
if (ARR_NDIM(array) < 1)
PG_RETURN_ARRAYTYPE_P(array);
if (dirstr != NULL)
{
if (!parse_sort_order(text_to_cstring(dirstr), &sort_asc, &nulls_first))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("second parameter must be a valid sort
direction")));
}
elmtyp = ARR_ELEMTYPE(array);
if (ARR_NDIM(array) > 1)
elmtyp = get_array_type(elmtyp);
typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
if (typentry == NULL || typentry->type_id != elmtyp)
{
typentry = lookup_type_cache(elmtyp, sort_asc ?
TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
if ((sort_asc && !OidIsValid(typentry->lt_opr)) ||
(!sort_asc && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify an ordering operator
for type %s",
format_type_be(elmtyp))));
fcinfo->flinfo->fn_extra = (void *) typentry;
}
---------------------------------------------------------------------
/*
* array_sort
*
* Sorts the array in either ascending or descending order.
* The array must be empty or one-dimensional.
*/
comments need to be updated.
typedef enum
PARSE_SORT_ORDER_DONE
} ParseSortOrderState;
last one, should have comma, like
"PARSE_SORT_ORDER_DONE, "
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 14:33 Junwang Zhao <[email protected]>
parent: jian he <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Junwang Zhao @ 2024-10-24 14:33 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Tom Lane <[email protected]>; Amit Langote <[email protected]>; David G. Johnston <[email protected]>; PostgreSQL Hackers <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Thu, Oct 24, 2024 at 8:40 PM jian he <[email protected]> wrote:
>
> On Wed, Oct 23, 2024 at 10:28 PM Junwang Zhao <[email protected]> wrote:
> > PFA v7 with multi-array support.
> >
>
> if (ARR_NDIM(array) == 1)
> {
> }
> else
> {
> }
> can be simplified.
> i think beginning part of array_sort can be like the following:
> (newline emitted)
>
> ---------------------------------------------------------------------
> if (ARR_NDIM(array) < 1)
> PG_RETURN_ARRAYTYPE_P(array);
> if (dirstr != NULL)
> {
> if (!parse_sort_order(text_to_cstring(dirstr), &sort_asc, &nulls_first))
> ereport(ERROR,
> (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
> errmsg("second parameter must be a valid sort
> direction")));
> }
> elmtyp = ARR_ELEMTYPE(array);
> if (ARR_NDIM(array) > 1)
> elmtyp = get_array_type(elmtyp);
I'm wondering should we cache the type entry for both element type and
the corresponding array type, for example if we have a table:
create table t(b int[]);
insert into t values ('{1,3}'),('{{2,3}}'),('{{1,2},{0,2}}');
with 1-d array and m-d array interleaved, then the following query will
call lookup_type_cache multiple times.
select array_sort((t.b)) from t;
> typentry = (TypeCacheEntry *) fcinfo->flinfo->fn_extra;
> if (typentry == NULL || typentry->type_id != elmtyp)
> {
> typentry = lookup_type_cache(elmtyp, sort_asc ?
> TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
> if ((sort_asc && !OidIsValid(typentry->lt_opr)) ||
> (!sort_asc && !OidIsValid(typentry->gt_opr)))
> ereport(ERROR,
> (errcode(ERRCODE_UNDEFINED_FUNCTION),
> errmsg("could not identify an ordering operator
> for type %s",
> format_type_be(elmtyp))));
> fcinfo->flinfo->fn_extra = (void *) typentry;
> }
> ---------------------------------------------------------------------
> /*
> * array_sort
> *
> * Sorts the array in either ascending or descending order.
> * The array must be empty or one-dimensional.
> */
> comments need to be updated.
will fix it in the next version of patch.
>
>
> typedef enum
> PARSE_SORT_ORDER_DONE
> } ParseSortOrderState;
>
> last one, should have comma, like
> "PARSE_SORT_ORDER_DONE, "
will fix it.
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 14:57 Aleksander Alekseev <[email protected]>
parent: Tom Lane <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: Aleksander Alekseev @ 2024-10-24 14:57 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Tom Lane <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; David G. Johnston <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi,
> It's hardly "general purpose" if it randomly refuses to
> sort certain types. I would say it should be able to sort
> anything that ORDER BY will handle --- and that certainly
> includes the cases shown here.
I wonder how useful / convenient the new function will be considering
that we already have CTEs and can do:
SELECT array_agg(x ORDER BY x) FROM unnest(ARRAY[5,1,3,2,4]) AS x;
Perhaps there are use cases I didn't consider?
--
Best regards,
Aleksander Alekseev
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 15:11 David G. Johnston <[email protected]>
parent: Aleksander Alekseev <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: David G. Johnston @ 2024-10-24 15:11 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Thu, Oct 24, 2024 at 7:58 AM Aleksander Alekseev <
[email protected]> wrote:
> Hi,
>
> > It's hardly "general purpose" if it randomly refuses to
> > sort certain types. I would say it should be able to sort
> > anything that ORDER BY will handle --- and that certainly
> > includes the cases shown here.
>
> I wonder how useful / convenient the new function will be considering
> that we already have CTEs and can do:
>
> SELECT array_agg(x ORDER BY x) FROM unnest(ARRAY[5,1,3,2,4]) AS x;
>
> Perhaps there are use cases I didn't consider?
>
>
Succinctness of expression. Plus I'm under the impression that a function
doing this is going to be somewhat faster than composing two functions
together within a multi-node subtree.
I feel like the same observation could have been made for array_shuffle but
we added that. This function being added feels to me like just completing
the set.
David J.
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 15:27 Aleksander Alekseev <[email protected]>
parent: David G. Johnston <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Aleksander Alekseev @ 2024-10-24 15:27 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: David G. Johnston <[email protected]>; Tom Lane <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi David,
>> > It's hardly "general purpose" if it randomly refuses to
>> > sort certain types. I would say it should be able to sort
>> > anything that ORDER BY will handle --- and that certainly
>> > includes the cases shown here.
>>
>> I wonder how useful / convenient the new function will be considering
>> that we already have CTEs and can do:
>>
>> SELECT array_agg(x ORDER BY x) FROM unnest(ARRAY[5,1,3,2,4]) AS x;
>>
>> Perhaps there are use cases I didn't consider?
>>
>
> Succinctness of expression. Plus I'm under the impression that a function doing this is going to be somewhat faster than composing two functions together within a multi-node subtree.
>
> I feel like the same observation could have been made for array_shuffle but we added that. This function being added feels to me like just completing the set.
Right. To clarify, I'm not opposed to array_sort(). In fact personally
I would prefer using it instead of array_agg() + unnest().
Just making an observation / thinking out loud that the requirement to
support everything ORDER BY handles / supports (and will handle /
support?) might make this function impractical to maintain.
array_shuffle() or a recently proposed array_reverse() [1] are rather
simple since they just move the array elements without looking at
them. array_sort() does look at the elements and thus is very
different.
Particularly, does the function really need to support dir => asc |
desc | asc nulls first | etc... ? Maybe array_sort() + array_reverse(
array_sort() ) will handle most of the practical cases? I don't know.
[1]: https://commitfest.postgresql.org/50/5314/
--
Best regards,
Aleksander Alekseev
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 15:40 David G. Johnston <[email protected]>
parent: Aleksander Alekseev <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: David G. Johnston @ 2024-10-24 15:40 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Thu, Oct 24, 2024 at 8:27 AM Aleksander Alekseev <
[email protected]> wrote:
>
> Just making an observation / thinking out loud that the requirement to
> support everything ORDER BY handles / supports (and will handle /
> support?) might make this function impractical to maintain.
>
> Particularly, does the function really need to support dir => asc |
> desc | asc nulls first | etc... ? Maybe array_sort() + array_reverse(
> array_sort() ) will handle most of the practical cases? I don't know.
>
> [1]: https://commitfest.postgresql.org/50/5314/
>
>
Composing function calls here seems quite undesirable from a performance
standpoint, but maybe I over-estimate the cost of
exploding-manipulating-freezing an array datum. Also, while I'm not in a
good position to judge the challenge of implementing the sort parameters I
would rather have them than not since the order by API has them (plus
performance). I also, maybe unreasonably, believe that our extensible type
system has already limited the maintenance burden.
David J.
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 16:06 Tom Lane <[email protected]>
parent: David G. Johnston <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Tom Lane @ 2024-10-24 16:06 UTC (permalink / raw)
To: David G. Johnston <[email protected]>; +Cc: Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
"David G. Johnston" <[email protected]> writes:
> Composing function calls here seems quite undesirable from a performance
> standpoint, but maybe I over-estimate the cost of
> exploding-manipulating-freezing an array datum. Also, while I'm not in a
> good position to judge the challenge of implementing the sort parameters I
> would rather have them than not since the order by API has them (plus
> performance). I also, maybe unreasonably, believe that our extensible type
> system has already limited the maintenance burden.
Oh! I had not actually looked at the details of what was being
proposed here. I imagined "array_sort(anyarray)" and the sort would
use the default sort order for the array's element type. This
business with a textual representation of a sort clause seems like
over-engineering ... except that it's also under-engineered, because
the parsing is lame and incomplete. (No USING option, and the fact
that collation comes from somewhere else seems impossibly confusing.)
Let's drop that. As already noted, if you need a non-default sort
order you can build what you want with a sub-select. The ambition
here should be to make easy cases easy.
regards, tom lane
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 16:47 David G. Johnston <[email protected]>
parent: Tom Lane <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: David G. Johnston @ 2024-10-24 16:47 UTC (permalink / raw)
To: Tom Lane <[email protected]>; +Cc: Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Thu, Oct 24, 2024 at 9:06 AM Tom Lane <[email protected]> wrote:
> This business with a textual representation of a sort clause seems like
> over-engineering ... except that it's also under-engineered, because
> the parsing is lame and incomplete. (No USING option, and the fact
> that collation comes from somewhere else seems impossibly confusing.)
> Let's drop that.
>
I can accept this outcome though an optional three-valued boolean sort
order (ascending and descending only) I'd argue is worth keeping. null
value placement too I guess, three-valued boolean (nulls_first).
David J.
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-24 17:18 Aleksander Alekseev <[email protected]>
parent: David G. Johnston <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Aleksander Alekseev @ 2024-10-24 17:18 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Junwang Zhao <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi,
> I can accept this outcome though an optional three-valued boolean sort order (ascending and descending only) I'd argue is worth keeping. null value placement too I guess, three-valued boolean (nulls_first).
Perhaps these optional arguments deserve separate discussions. I
suggest merging something everyone agrees on first. This will simplify
the review process and allow us to deliver value to the users quickly.
Arguments like `reverse => true` and `nulls_first => true` can always
be implemented and added as separate patches.
--
Best regards,
Aleksander Alekseev
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v23 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-10-25 03:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-10-25 03:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index efa730c167..a80263f90d 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4c97690908..a6d2f7f4ba 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3823,3 +3834,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index ef0b560f5e..b30b9f2675 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Oct_25_13_04_53_2024_648)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v23-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v23 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-10-25 03:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-10-25 03:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index efa730c167..a80263f90d 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4c97690908..a6d2f7f4ba 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3823,3 +3834,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index ef0b560f5e..b30b9f2675 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Oct_25_13_04_53_2024_648)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v23-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v23 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-10-25 03:56 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-10-25 03:56 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index efa730c167..a80263f90d 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 4c97690908..a6d2f7f4ba 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -98,7 +98,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2956,6 +2963,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3823,3 +3834,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index ef0b560f5e..b30b9f2675 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -577,6 +577,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1860,6 +1861,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3199,6 +3203,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Fri_Oct_25_13_04_53_2024_648)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v23-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-25 12:02 Junwang Zhao <[email protected]>
parent: Aleksander Alekseev <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-25 12:02 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Fri, Oct 25, 2024 at 1:19 AM Aleksander Alekseev
<[email protected]> wrote:
>
> Hi,
>
> > I can accept this outcome though an optional three-valued boolean sort order (ascending and descending only) I'd argue is worth keeping. null value placement too I guess, three-valued boolean (nulls_first).
>
> Perhaps these optional arguments deserve separate discussions. I
> suggest merging something everyone agrees on first. This will simplify
> the review process and allow us to deliver value to the users quickly.
> Arguments like `reverse => true` and `nulls_first => true` can always
> be implemented and added as separate patches.
As this patch uses the tuplesort infrastructure, we need to supply the
sortOperator, sortCollation and nullsFirstFlag, I tend to agree with
David. I admit that the parsing part is not good, so I will remove it
by using two boolean parameters Jian suggested earlier.
Will send out another version by tomorrow.
>
> --
> Best regards,
> Aleksander Alekseev
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-26 04:20 Junwang Zhao <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-26 04:20 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Fri, Oct 25, 2024 at 8:02 PM Junwang Zhao <[email protected]> wrote:
>
> On Fri, Oct 25, 2024 at 1:19 AM Aleksander Alekseev
> <[email protected]> wrote:
> >
> > Hi,
> >
> > > I can accept this outcome though an optional three-valued boolean sort order (ascending and descending only) I'd argue is worth keeping. null value placement too I guess, three-valued boolean (nulls_first).
> >
> > Perhaps these optional arguments deserve separate discussions. I
> > suggest merging something everyone agrees on first. This will simplify
> > the review process and allow us to deliver value to the users quickly.
> > Arguments like `reverse => true` and `nulls_first => true` can always
> > be implemented and added as separate patches.
>
> As this patch uses the tuplesort infrastructure, we need to supply the
> sortOperator, sortCollation and nullsFirstFlag, I tend to agree with
> David. I admit that the parsing part is not good, so I will remove it
> by using two boolean parameters Jian suggested earlier.
>
> Will send out another version by tomorrow.
Based on the previous discussion, I split it into two patches in V8.
0001 is the general sort part without `is_ascending` or `nulls_first`,
the sort order is determined by the "<" operator of the element type.
It also cached the type entry of both eletyp and the corresponding
array type.
0002 adds the `is_ascending` and `nulls_first` part, it now uses
two boolean parameters instead of parsing one text parameter.
>
> >
> > --
> > Best regards,
> > Aleksander Alekseev
>
>
>
> --
> Regards
> Junwang Zhao
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v8-0002-support-sort-order-and-nullsfirst-flag.patch (8.4K, ../../CAEG8a3KztaRQT9X8UfHACJ1Qszw93guLpGJG0pRHS6-OYoCR3Q@mail.gmail.com/2-v8-0002-support-sort-order-and-nullsfirst-flag.patch)
download | inline diff:
From 5a89f51da4ce7adc5d26efab19b199507300dd0f Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Sat, 26 Oct 2024 03:53:17 +0000
Subject: [PATCH v8 2/2] support sort order and nullsfirst flag
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 11 +++++--
src/backend/utils/adt/array_userfuncs.c | 42 ++++++++++++++++++++-----
src/include/catalog/pg_proc.dat | 6 ++++
src/test/regress/expected/arrays.out | 26 +++++++++++++++
src/test/regress/sql/arrays.sql | 8 +++++
5 files changed, 84 insertions(+), 9 deletions(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 668fe5d358..129699aa8a 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20426,12 +20426,19 @@ SELECT NULLIF(value, '(none)') ...
<indexterm>
<primary>array_sort</primary>
</indexterm>
- <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional>)
+ <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional> <optional>, <parameter>is_ascending</parameter> <type>boolean</type> <optional>, <parameter>nulls_first</parameter> <type>boolean</type></optional></optional>)
<returnvalue>anyarray</returnvalue>
</para>
<para>
- Sorts the first dimension of the array. If the <literal>COLLATE</literal> option is specified then sorting is based on <replaceable>collation_name</replaceable>, otherwise
+ Sorts the first dimension of the array.
+ </para>
+ <para>
+ If the <literal>COLLATE</literal> option is specified then sorting is based on <replaceable>collation_name</replaceable>, otherwise
using array element type's collation.
+ If <parameter>is_ascending</parameter> is true then sort by ascending order, otherwise descending order.
+ <parameter>is_ascending</parameter> defaults to true.
+ If <parameter>nulls_first</parameter> is true then nulls appear before non-null values, otherwise nulls appear after non-null values.
+ <parameter>nulls_first</parameter> defaults to the opposite of <parameter>is_ascending</parameter> if not provided.
</para>
<para>
<literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 9b775313ca..1ff61033b2 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1698,7 +1698,6 @@ typedef struct ArraySortCachedInfo
* array_sort
*
* Sorts the first dimension of the array.
- * The sort order is determined by the "<" operator of the element type.
*/
Datum
array_sort(PG_FUNCTION_ARGS)
@@ -1714,10 +1713,23 @@ array_sort(PG_FUNCTION_ARGS)
Datum value;
bool isnull;
ArrayBuildStateAny *astate = NULL;
+ bool is_ascending = true;
+ bool nulls_first = false;
if (ARR_NDIM(array) < 1)
PG_RETURN_ARRAYTYPE_P(array);
+ if (PG_NARGS() > 1)
+ {
+ is_ascending = PG_GETARG_BOOL(1);
+
+ /*
+ * If nulls_first not provided, it defaults to the opposite of
+ * is_ascending.
+ */
+ nulls_first = PG_NARGS() > 2 ? PG_GETARG_BOOL(2) : !is_ascending;
+ }
+
elmtyp = ARR_ELEMTYPE(array);
cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
if (cache_info == NULL)
@@ -1735,8 +1747,10 @@ array_sort(PG_FUNCTION_ARGS)
typentry = cache_info->typentry;
if (typentry == NULL || typentry->type_id != elmtyp)
{
- typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(elmtyp,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1756,8 +1770,10 @@ array_sort(PG_FUNCTION_ARGS)
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
format_type_be(elmtyp))));
- typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(array_type,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1767,9 +1783,9 @@ array_sort(PG_FUNCTION_ARGS)
}
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
- typentry->lt_opr,
+ is_ascending ? typentry->lt_opr : typentry->gt_opr,
collation,
- false, work_mem, NULL, false);
+ nulls_first, work_mem, NULL, false);
array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
while (array_iterate(array_iterator, &value, &isnull))
@@ -1795,3 +1811,15 @@ array_sort(PG_FUNCTION_ARGS)
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
+
+Datum
+array_sort_order_nulls_first(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 79641da6ff..9bbc97445c 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,12 @@
{ oid => '8810', descr => 'sort array',
proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort_order'},
+{ oid => '8812', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool bool', prosrc => 'array_sort_order_nulls_first'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index ba1da71da1..f4781233e0 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2728,6 +2728,32 @@ SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo}
(1 row)
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+ array_sort
+--------------------------------
+ {NULL,1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+ array_sort
+--------------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1,NULL}
+(1 row)
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
array_sort
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index b2a814d195..b98e111329 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -834,5 +834,13 @@ SELECT array_sort('{1,3,5,2,4,6}'::int[]);
SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
--
2.39.5
[application/octet-stream] v8-0001-general-purpose-array_sort.patch (10.3K, ../../CAEG8a3KztaRQT9X8UfHACJ1Qszw93guLpGJG0pRHS6-OYoCR3Q@mail.gmail.com/3-v8-0001-general-purpose-array_sort.patch)
download | inline diff:
From dd5a1cdb82212374da50cc52ce69c88544011356 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v8 1/2] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 110 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 3 +
src/test/regress/expected/arrays.out | 32 +++++
.../regress/expected/collate.icu.utf8.out | 13 +++
src/test/regress/sql/arrays.sql | 9 ++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 190 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 7be0324ac8..668fe5d358 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20421,6 +20421,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional>)
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. If the <literal>COLLATE</literal> option is specified then sorting is based on <replaceable>collation_name</replaceable>, otherwise
+ using array element type's collation.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 6599be2ec5..9b775313ca 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1685,3 +1687,111 @@ array_sample(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry;
+ TypeCacheEntry *array_typentry;
+} ArraySortCachedInfo;
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ARR_NDIM(array) == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 1ec0d6f6b5..79641da6ff 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1741,6 +1741,9 @@
{ oid => '6216', descr => 'take samples from array',
proname => 'array_sample', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray int4', prosrc => 'array_sample' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index a6d81fd5f9..ba1da71da1 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2703,3 +2703,35 @@ SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
ERROR: sample size must be between 0 and 6
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
ERROR: sample size must be between 0 and 6
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 47058dfde5..b2a814d195 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -827,3 +827,12 @@ SELECT array_dims(array_sample('[-1:2][2:3]={{1,2},{3,NULL},{5,6},{7,8}}'::int[]
SELECT array_dims(array_sample('{{{1,2},{3,NULL}},{{5,6},{7,8}},{{9,10},{11,12}}}'::int[], 2));
SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 171a7dd5d2..eceba9605e 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-28 16:48 Aleksander Alekseev <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 2 replies; 166+ messages in thread
From: Aleksander Alekseev @ 2024-10-28 16:48 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Junwang Zhao <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi,
> Based on the previous discussion, I split it into two patches in V8.
>
> 0001 is the general sort part without `is_ascending` or `nulls_first`,
> the sort order is determined by the "<" operator of the element type.
> It also cached the type entry of both eletyp and the corresponding
> array type.
>
> 0002 adds the `is_ascending` and `nulls_first` part, it now uses
> two boolean parameters instead of parsing one text parameter.
Thanks for the update patch set. Here are some comments.
0001:
> +{ oid => '8810', descr => 'sort array',
> + proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
> + proargtypes => 'anyarray', prosrc => 'array_sort'},
I would expect that array_sort() should be IMMUTABLE. Is there a
reason for it to be VOLATILE?
> + <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional>)
> + <returnvalue>anyarray</returnvalue>
It seems to me that the part about using COLLATE should be moved
below, to the description / examples section, since it's not part of
the function signature.
Also the description should be more specific about how NULLs are
sorted. NULLs also should be covered by tests.
0002:
> <parameter>is_ascending</parameter>
I really believe this name is not the best one. I suggest using
`reverse => true`. `nulls_first` is OK.
> +Datum
> +array_sort_order(PG_FUNCTION_ARGS)
> +{
> + return array_sort(fcinfo);
> +}
> +
> +Datum
> +array_sort_order_nulls_first(PG_FUNCTION_ARGS)
> +{
> + return array_sort(fcinfo);
> +}
Any reason not to specify array_sort in pg_proc.dat?
The tests cover is_ascending => true | false, which is OK, but only
(is_ascending = true, nulls_first => true) and (is_ascending => false,
nulls_fist => false). For the case when both optional arguments are
specified you have to test at least 4 combinations.
--
Best regards,
Aleksander Alekseev
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-29 08:28 jian he <[email protected]>
parent: Aleksander Alekseev <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: jian he @ 2024-10-29 08:28 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Junwang Zhao <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Tue, Oct 29, 2024 at 12:48 AM Aleksander Alekseev
<[email protected]> wrote:.
>
> 0001:
>
> > +{ oid => '8810', descr => 'sort array',
> > + proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
> > + proargtypes => 'anyarray', prosrc => 'array_sort'},
>
> I would expect that array_sort() should be IMMUTABLE. Is there a
> reason for it to be VOLATILE?
>
https://www.postgresql.org/docs/current/sql-createfunction.html says:
IMMUTABLE indicates that the function cannot modify the database and always
returns the same result when given the same argument values; that is, it does
not do database lookups or otherwise use information not directly present in its
argument list. If this option is given, any call of the function with
all-constant arguments can be immediately replaced with the function value.
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
This error can happen. I think this conflicts with the doc IMMUTABLE
description.
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-29 12:14 Aleksander Alekseev <[email protected]>
parent: jian he <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Aleksander Alekseev @ 2024-10-29 12:14 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: jian he <[email protected]>; Junwang Zhao <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Jian,
> IMMUTABLE indicates that the function cannot modify the database and always
> returns the same result when given the same argument values; that is, it does
> not do database lookups or otherwise use information not directly present in its
> argument list. If this option is given, any call of the function with
> all-constant arguments can be immediately replaced with the function value.
>
>
> + {
> + typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
> + if (!OidIsValid(typentry->lt_opr))
> + ereport(ERROR,
> + (errcode(ERRCODE_UNDEFINED_FUNCTION),
> + errmsg("could not identify ordering operator for type %s",
> + format_type_be(elmtyp))));
>
> This error can happen. I think this conflicts with the doc IMMUTABLE
> description.
lookup_type_cache() is used at least by array_position() which is
marked as IMMUTABLE, so I believe this is fine. Similarly functions
dealing with timezones can return different results between the DBMS
restarts / updates, but we don't care and mark them IMMUTABLE anyway.
Otherwise we couldn't use these functions in functional indexes which
will make them rather useless.
--
Best regards,
Aleksander Alekseev
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-29 14:49 Junwang Zhao <[email protected]>
parent: Aleksander Alekseev <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-29 14:49 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Aleksander,
On Tue, Oct 29, 2024 at 12:48 AM Aleksander Alekseev
<[email protected]> wrote:
>
> Hi,
>
> > Based on the previous discussion, I split it into two patches in V8.
> >
> > 0001 is the general sort part without `is_ascending` or `nulls_first`,
> > the sort order is determined by the "<" operator of the element type.
> > It also cached the type entry of both eletyp and the corresponding
> > array type.
> >
> > 0002 adds the `is_ascending` and `nulls_first` part, it now uses
> > two boolean parameters instead of parsing one text parameter.
>
> Thanks for the update patch set. Here are some comments.
>
> 0001:
>
> > +{ oid => '8810', descr => 'sort array',
> > + proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
> > + proargtypes => 'anyarray', prosrc => 'array_sort'},
>
> I would expect that array_sort() should be IMMUTABLE. Is there a
> reason for it to be VOLATILE?
I saw Jian's reply about this, but I tend to agree with you, so remove
provolatile => 'v'.
>
> > + <function>array_sort</function> ( <type>anyarray</type> <optional> COLLATE <replaceable>collation_name</replaceable> </optional>)
> > + <returnvalue>anyarray</returnvalue>
>
> It seems to me that the part about using COLLATE should be moved
> below, to the description / examples section, since it's not part of
> the function signature.
Agree, fixed with my own words, help needed with the wording.
>
> Also the description should be more specific about how NULLs are
> sorted. NULLs also should be covered by tests.
Fixed.
>
> 0002:
>
> > <parameter>is_ascending</parameter>
>
> I really believe this name is not the best one. I suggest using
> `reverse => true`. `nulls_first` is OK.
Not sure about this, I think `is_ascending` has a more precise
meaning, while `reverse` doesn't show any hint about ascending or
descending, just keep it right now, let's see others' opinions.
>
> > +Datum
> > +array_sort_order(PG_FUNCTION_ARGS)
> > +{
> > + return array_sort(fcinfo);
> > +}
> > +
> > +Datum
> > +array_sort_order_nulls_first(PG_FUNCTION_ARGS)
> > +{
> > + return array_sort(fcinfo);
> > +}
>
> Any reason not to specify array_sort in pg_proc.dat?
It is specified in 0001 (see oid => '8810').
>
> The tests cover is_ascending => true | false, which is OK, but only
> (is_ascending = true, nulls_first => true) and (is_ascending => false,
> nulls_fist => false). For the case when both optional arguments are
> specified you have to test at least 4 combinations.
The omitted two is the same as the two with two parameters specified,
anyway, add all 4 cases in v9.
>
> --
> Best regards,
> Aleksander Alekseev
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v9-0002-support-sort-order-and-nullsfirst-flag.patch (8.6K, ../../CAEG8a3JwU4N+2uk6=iJHex24vEAyTmGZKRr9aWK4BfSuKjQBMQ@mail.gmail.com/2-v9-0002-support-sort-order-and-nullsfirst-flag.patch)
download | inline diff:
From a2ef7b07f26edeeac3a914f42a13a831e3df4419 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Sat, 26 Oct 2024 03:53:17 +0000
Subject: [PATCH v9 2/2] support sort order and nullsfirst flag
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 6 +++-
src/backend/utils/adt/array_userfuncs.c | 42 ++++++++++++++++++++-----
src/include/catalog/pg_proc.dat | 6 ++++
src/test/regress/expected/arrays.out | 38 ++++++++++++++++++++++
src/test/regress/sql/arrays.sql | 10 ++++++
5 files changed, 94 insertions(+), 8 deletions(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 4095e23567..abde23ffc3 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20426,12 +20426,16 @@ SELECT NULLIF(value, '(none)') ...
<indexterm>
<primary>array_sort</primary>
</indexterm>
- <function>array_sort</function> ( <type>anyarray</type> )
+ <function>array_sort</function> ( <type>anyarray</type> <optional>, <parameter>is_ascending</parameter> <type>boolean</type> <optional>, <parameter>nulls_first</parameter> <type>boolean</type></optional></optional>)
<returnvalue>anyarray</returnvalue>
</para>
<para>
Sorts the first dimension of the array. The sort order is determined by the "<" operator of the element type, nulls will appear after non-null values.
The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ If <parameter>is_ascending</parameter> is true then sort by ascending order, otherwise descending order.
+ <parameter>is_ascending</parameter> defaults to true.
+ If <parameter>nulls_first</parameter> is true then nulls appear before non-null values, otherwise nulls appear after non-null values.
+ <parameter>nulls_first</parameter> defaults to the opposite of <parameter>is_ascending</parameter> if not provided.
</para>
<para>
<literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 9b775313ca..1ff61033b2 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1698,7 +1698,6 @@ typedef struct ArraySortCachedInfo
* array_sort
*
* Sorts the first dimension of the array.
- * The sort order is determined by the "<" operator of the element type.
*/
Datum
array_sort(PG_FUNCTION_ARGS)
@@ -1714,10 +1713,23 @@ array_sort(PG_FUNCTION_ARGS)
Datum value;
bool isnull;
ArrayBuildStateAny *astate = NULL;
+ bool is_ascending = true;
+ bool nulls_first = false;
if (ARR_NDIM(array) < 1)
PG_RETURN_ARRAYTYPE_P(array);
+ if (PG_NARGS() > 1)
+ {
+ is_ascending = PG_GETARG_BOOL(1);
+
+ /*
+ * If nulls_first not provided, it defaults to the opposite of
+ * is_ascending.
+ */
+ nulls_first = PG_NARGS() > 2 ? PG_GETARG_BOOL(2) : !is_ascending;
+ }
+
elmtyp = ARR_ELEMTYPE(array);
cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
if (cache_info == NULL)
@@ -1735,8 +1747,10 @@ array_sort(PG_FUNCTION_ARGS)
typentry = cache_info->typentry;
if (typentry == NULL || typentry->type_id != elmtyp)
{
- typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(elmtyp,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1756,8 +1770,10 @@ array_sort(PG_FUNCTION_ARGS)
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
format_type_be(elmtyp))));
- typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(array_type,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1767,9 +1783,9 @@ array_sort(PG_FUNCTION_ARGS)
}
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
- typentry->lt_opr,
+ is_ascending ? typentry->lt_opr : typentry->gt_opr,
collation,
- false, work_mem, NULL, false);
+ nulls_first, work_mem, NULL, false);
array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
while (array_iterate(array_iterator, &value, &isnull))
@@ -1795,3 +1811,15 @@ array_sort(PG_FUNCTION_ARGS)
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
+
+Datum
+array_sort_order_nulls_first(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 5609368772..1987ef249e 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,12 @@
{ oid => '8810', descr => 'sort array',
proname => 'array_sort', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort_order'},
+{ oid => '8812', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool bool', prosrc => 'array_sort_order_nulls_first'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index da5ffe5eba..d4fa66dbf5 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2734,6 +2734,44 @@ SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo,NULL}
(1 row)
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+ array_sort
+--------------------------------
+ {NULL,1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+ array_sort
+--------------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1,NULL}
+(1 row)
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
array_sort
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index b4cfb0b038..768332641b 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -835,5 +835,15 @@ SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
--
2.39.5
[application/octet-stream] v9-0001-general-purpose-array_sort.patch (10.5K, ../../CAEG8a3JwU4N+2uk6=iJHex24vEAyTmGZKRr9aWK4BfSuKjQBMQ@mail.gmail.com/3-v9-0001-general-purpose-array_sort.patch)
download | inline diff:
From 2f47117bd4d16782c9237b7459a1a535ec7f9ff9 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v9 1/2] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 110 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 3 +
src/test/regress/expected/arrays.out | 38 ++++++
.../regress/expected/collate.icu.utf8.out | 13 +++
src/test/regress/sql/arrays.sql | 10 ++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 197 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 7be0324ac8..4095e23567 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20421,6 +20421,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> )
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. The sort order is determined by the "<" operator of the element type, nulls will appear after non-null values.
+ The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 6599be2ec5..9b775313ca 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1685,3 +1687,111 @@ array_sample(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry;
+ TypeCacheEntry *array_typentry;
+} ArraySortCachedInfo;
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ARR_NDIM(array) == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 1ec0d6f6b5..5609368772 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1741,6 +1741,9 @@
{ oid => '6216', descr => 'take samples from array',
proname => 'array_sample', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray int4', prosrc => 'array_sample' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index a6d81fd5f9..da5ffe5eba 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2703,3 +2703,41 @@ SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
ERROR: sample size must be between 0 and 6
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
ERROR: sample size must be between 0 and 6
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 47058dfde5..b4cfb0b038 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -827,3 +827,13 @@ SELECT array_dims(array_sample('[-1:2][2:3]={{1,2},{3,NULL},{5,6},{7,8}}'::int[]
SELECT array_dims(array_sample('{{{1,2},{3,NULL}},{{5,6},{7,8}},{{9,10},{11,12}}}'::int[], 2));
SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 171a7dd5d2..eceba9605e 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-30 13:29 Aleksander Alekseev <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Aleksander Alekseev @ 2024-10-30 13:29 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Junwang Zhao <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi,
Thanks for the updated patch set.
> > > +Datum
> > > +array_sort_order(PG_FUNCTION_ARGS)
> > > +{
> > > + return array_sort(fcinfo);
> > > +}
> > > +
> > > +Datum
> > > +array_sort_order_nulls_first(PG_FUNCTION_ARGS)
> > > +{
> > > + return array_sort(fcinfo);
> > > +}
> >
> > Any reason not to specify array_sort in pg_proc.dat?
>
> It is specified in 0001 (see oid => '8810').
What I meant was that I don't think these wrapper functions are
needed. I think you can just do:
```
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort'}, <--
array_sort is used directly in `prosrc`
```
... unless I'm missing something.
--
Best regards,
Aleksander Alekseev
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-30 14:17 Junwang Zhao <[email protected]>
parent: Aleksander Alekseev <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-30 14:17 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi,
On Wed, Oct 30, 2024 at 9:29 PM Aleksander Alekseev
<[email protected]> wrote:
>
> Hi,
>
> Thanks for the updated patch set.
>
> > > > +Datum
> > > > +array_sort_order(PG_FUNCTION_ARGS)
> > > > +{
> > > > + return array_sort(fcinfo);
> > > > +}
> > > > +
> > > > +Datum
> > > > +array_sort_order_nulls_first(PG_FUNCTION_ARGS)
> > > > +{
> > > > + return array_sort(fcinfo);
> > > > +}
> > >
> > > Any reason not to specify array_sort in pg_proc.dat?
> >
> > It is specified in 0001 (see oid => '8810').
>
> What I meant was that I don't think these wrapper functions are
> needed. I think you can just do:
>
> ```
> +{ oid => '8811', descr => 'sort array',
> + proname => 'array_sort', prorettype => 'anyarray',
> + proargtypes => 'anyarray bool', prosrc => 'array_sort'}, <--
> array_sort is used directly in `prosrc`
> ```
>
> ... unless I'm missing something.
There is a opr sanity check for this[1], if we remove these wrapper functions,
regression test will fail with:
- oid | proname | oid | proname
------+---------+-----+---------
-(0 rows)
+ oid | proname | oid | proname
+------+------------+------+------------
+ 8811 | array_sort | 8812 | array_sort
+ 8810 | array_sort | 8811 | array_sort
+ 8810 | array_sort | 8812 | array_sort
+(3 rows)
[1]:
-- Considering only built-in procs (prolang = 12), look for multiple uses
-- of the same internal function (ie, matching prosrc fields). It's OK to
-- have several entries with different pronames for the same internal function,
-- but conflicts in the number of arguments and other critical items should
-- be complained of. (We don't check data types here; see next query.)
-- Note: ignore aggregate functions here, since they all point to the same
-- dummy built-in function.
SELECT p1.oid, p1.proname, p2.oid, p2.proname
FROM pg_proc AS p1, pg_proc AS p2
WHERE p1.oid < p2.oid AND
p1.prosrc = p2.prosrc AND
p1.prolang = 12 AND p2.prolang = 12 AND
(p1.prokind != 'a' OR p2.prokind != 'a') AND
(p1.prolang != p2.prolang OR
p1.prokind != p2.prokind OR
p1.prosecdef != p2.prosecdef OR
p1.proleakproof != p2.proleakproof OR
p1.proisstrict != p2.proisstrict OR
p1.proretset != p2.proretset OR
p1.provolatile != p2.provolatile OR
p1.pronargs != p2.pronargs);
>
> --
> Best regards,
> Aleksander Alekseev
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-10-30 14:41 Junwang Zhao <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-10-30 14:41 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Wed, Oct 30, 2024 at 10:17 PM Junwang Zhao <[email protected]> wrote:
>
> Hi,
>
> On Wed, Oct 30, 2024 at 9:29 PM Aleksander Alekseev
> <[email protected]> wrote:
> >
> > Hi,
> >
> > Thanks for the updated patch set.
> >
> > > > > +Datum
> > > > > +array_sort_order(PG_FUNCTION_ARGS)
> > > > > +{
> > > > > + return array_sort(fcinfo);
> > > > > +}
> > > > > +
> > > > > +Datum
> > > > > +array_sort_order_nulls_first(PG_FUNCTION_ARGS)
> > > > > +{
> > > > > + return array_sort(fcinfo);
> > > > > +}
> > > >
> > > > Any reason not to specify array_sort in pg_proc.dat?
> > >
> > > It is specified in 0001 (see oid => '8810').
> >
> > What I meant was that I don't think these wrapper functions are
> > needed. I think you can just do:
> >
> > ```
> > +{ oid => '8811', descr => 'sort array',
> > + proname => 'array_sort', prorettype => 'anyarray',
> > + proargtypes => 'anyarray bool', prosrc => 'array_sort'}, <--
> > array_sort is used directly in `prosrc`
> > ```
> >
> > ... unless I'm missing something.
>
> There is a opr sanity check for this[1], if we remove these wrapper functions,
> regression test will fail with:
>
> - oid | proname | oid | proname
> ------+---------+-----+---------
> -(0 rows)
> + oid | proname | oid | proname
> +------+------------+------+------------
> + 8811 | array_sort | 8812 | array_sort
> + 8810 | array_sort | 8811 | array_sort
> + 8810 | array_sort | 8812 | array_sort
> +(3 rows)
>
>
> [1]:
>
> -- Considering only built-in procs (prolang = 12), look for multiple uses
> -- of the same internal function (ie, matching prosrc fields). It's OK to
> -- have several entries with different pronames for the same internal function,
> -- but conflicts in the number of arguments and other critical items should
> -- be complained of. (We don't check data types here; see next query.)
> -- Note: ignore aggregate functions here, since they all point to the same
> -- dummy built-in function.
>
> SELECT p1.oid, p1.proname, p2.oid, p2.proname
> FROM pg_proc AS p1, pg_proc AS p2
> WHERE p1.oid < p2.oid AND
> p1.prosrc = p2.prosrc AND
> p1.prolang = 12 AND p2.prolang = 12 AND
> (p1.prokind != 'a' OR p2.prokind != 'a') AND
> (p1.prolang != p2.prolang OR
> p1.prokind != p2.prokind OR
> p1.prosecdef != p2.prosecdef OR
> p1.proleakproof != p2.proleakproof OR
> p1.proisstrict != p2.proisstrict OR
> p1.proretset != p2.proretset OR
> p1.provolatile != p2.provolatile OR
> p1.pronargs != p2.pronargs);
>
> >
> > --
> > Best regards,
> > Aleksander Alekseev
>
>
>
> --
> Regards
> Junwang Zhao
CFbot failed with doc build, v10 fixed that.
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v10-0002-support-sort-order-and-nullsfirst-flag.patch (8.7K, ../../CAEG8a3Lzxv=QVt0Q8Mia2iFZUWvwAWdCk8btHp9GiYgTMK_+ig@mail.gmail.com/2-v10-0002-support-sort-order-and-nullsfirst-flag.patch)
download | inline diff:
From 1710c22a854b103562df3ff91a038e3fffd74ba3 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Sat, 26 Oct 2024 03:53:17 +0000
Subject: [PATCH v10 2/2] support sort order and nullsfirst flag
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 6 +++-
src/backend/utils/adt/array_userfuncs.c | 42 ++++++++++++++++++++-----
src/include/catalog/pg_proc.dat | 6 ++++
src/test/regress/expected/arrays.out | 38 ++++++++++++++++++++++
src/test/regress/sql/arrays.sql | 10 ++++++
5 files changed, 94 insertions(+), 8 deletions(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index bfe6b6a4de..f0133fe774 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20426,12 +20426,16 @@ SELECT NULLIF(value, '(none)') ...
<indexterm>
<primary>array_sort</primary>
</indexterm>
- <function>array_sort</function> ( <type>anyarray</type> )
+ <function>array_sort</function> ( <type>anyarray</type> <optional>, <parameter>is_ascending</parameter> <type>boolean</type> <optional>, <parameter>nulls_first</parameter> <type>boolean</type></optional></optional>)
<returnvalue>anyarray</returnvalue>
</para>
<para>
Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ If <parameter>is_ascending</parameter> is true then sort by ascending order, otherwise descending order.
+ <parameter>is_ascending</parameter> defaults to true.
+ If <parameter>nulls_first</parameter> is true then nulls appear before non-null values, otherwise nulls appear after non-null values.
+ <parameter>nulls_first</parameter> defaults to the opposite of <parameter>is_ascending</parameter> if not provided.
</para>
<para>
<literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 9b775313ca..1ff61033b2 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1698,7 +1698,6 @@ typedef struct ArraySortCachedInfo
* array_sort
*
* Sorts the first dimension of the array.
- * The sort order is determined by the "<" operator of the element type.
*/
Datum
array_sort(PG_FUNCTION_ARGS)
@@ -1714,10 +1713,23 @@ array_sort(PG_FUNCTION_ARGS)
Datum value;
bool isnull;
ArrayBuildStateAny *astate = NULL;
+ bool is_ascending = true;
+ bool nulls_first = false;
if (ARR_NDIM(array) < 1)
PG_RETURN_ARRAYTYPE_P(array);
+ if (PG_NARGS() > 1)
+ {
+ is_ascending = PG_GETARG_BOOL(1);
+
+ /*
+ * If nulls_first not provided, it defaults to the opposite of
+ * is_ascending.
+ */
+ nulls_first = PG_NARGS() > 2 ? PG_GETARG_BOOL(2) : !is_ascending;
+ }
+
elmtyp = ARR_ELEMTYPE(array);
cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
if (cache_info == NULL)
@@ -1735,8 +1747,10 @@ array_sort(PG_FUNCTION_ARGS)
typentry = cache_info->typentry;
if (typentry == NULL || typentry->type_id != elmtyp)
{
- typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(elmtyp,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1756,8 +1770,10 @@ array_sort(PG_FUNCTION_ARGS)
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
format_type_be(elmtyp))));
- typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(array_type,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1767,9 +1783,9 @@ array_sort(PG_FUNCTION_ARGS)
}
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
- typentry->lt_opr,
+ is_ascending ? typentry->lt_opr : typentry->gt_opr,
collation,
- false, work_mem, NULL, false);
+ nulls_first, work_mem, NULL, false);
array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
while (array_iterate(array_iterator, &value, &isnull))
@@ -1795,3 +1811,15 @@ array_sort(PG_FUNCTION_ARGS)
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
+
+Datum
+array_sort_order_nulls_first(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 5609368772..1987ef249e 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,12 @@
{ oid => '8810', descr => 'sort array',
proname => 'array_sort', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort_order'},
+{ oid => '8812', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool bool', prosrc => 'array_sort_order_nulls_first'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index da5ffe5eba..d4fa66dbf5 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2734,6 +2734,44 @@ SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo,NULL}
(1 row)
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+ array_sort
+--------------------------------
+ {NULL,1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+ array_sort
+--------------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1,NULL}
+(1 row)
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
array_sort
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index b4cfb0b038..768332641b 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -835,5 +835,15 @@ SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
--
2.39.5
[application/octet-stream] v10-0001-general-purpose-array_sort.patch (10.5K, ../../CAEG8a3Lzxv=QVt0Q8Mia2iFZUWvwAWdCk8btHp9GiYgTMK_+ig@mail.gmail.com/3-v10-0001-general-purpose-array_sort.patch)
download | inline diff:
From d65f0cd33185dab1a663d0a344872bc0196524a0 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v10 1/2] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 110 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 3 +
src/test/regress/expected/arrays.out | 38 ++++++
.../regress/expected/collate.icu.utf8.out | 13 +++
src/test/regress/sql/arrays.sql | 10 ++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 197 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 7be0324ac8..bfe6b6a4de 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20421,6 +20421,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> )
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
+ The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 6599be2ec5..9b775313ca 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1685,3 +1687,111 @@ array_sample(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry;
+ TypeCacheEntry *array_typentry;
+} ArraySortCachedInfo;
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ARR_NDIM(array) == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 1ec0d6f6b5..5609368772 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1741,6 +1741,9 @@
{ oid => '6216', descr => 'take samples from array',
proname => 'array_sample', provolatile => 'v', prorettype => 'anyarray',
proargtypes => 'anyarray int4', prosrc => 'array_sample' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index a6d81fd5f9..da5ffe5eba 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2703,3 +2703,41 @@ SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
ERROR: sample size must be between 0 and 6
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
ERROR: sample size must be between 0 and 6
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 47058dfde5..b4cfb0b038 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -827,3 +827,13 @@ SELECT array_dims(array_sample('[-1:2][2:3]={{1,2},{3,NULL},{5,6},{7,8}}'::int[]
SELECT array_dims(array_sample('{{{1,2},{3,NULL}},{{5,6},{7,8}},{{9,10},{11,12}}}'::int[], 2));
SELECT array_sample('{1,2,3,4,5,6}'::int[], -1); -- fail
SELECT array_sample('{1,2,3,4,5,6}'::int[], 7); --fail
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 171a7dd5d2..eceba9605e 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-03 03:33 Junwang Zhao <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 2 replies; 166+ messages in thread
From: Junwang Zhao @ 2024-11-03 03:33 UTC (permalink / raw)
To: Aleksander Alekseev <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Wed, Oct 30, 2024 at 10:41 PM Junwang Zhao <[email protected]> wrote:
>
> On Wed, Oct 30, 2024 at 10:17 PM Junwang Zhao <[email protected]> wrote:
> >
> > Hi,
> >
> > On Wed, Oct 30, 2024 at 9:29 PM Aleksander Alekseev
> > <[email protected]> wrote:
> > >
> > > Hi,
> > >
> > > Thanks for the updated patch set.
> > >
> > > > > > +Datum
> > > > > > +array_sort_order(PG_FUNCTION_ARGS)
> > > > > > +{
> > > > > > + return array_sort(fcinfo);
> > > > > > +}
> > > > > > +
> > > > > > +Datum
> > > > > > +array_sort_order_nulls_first(PG_FUNCTION_ARGS)
> > > > > > +{
> > > > > > + return array_sort(fcinfo);
> > > > > > +}
> > > > >
> > > > > Any reason not to specify array_sort in pg_proc.dat?
> > > >
> > > > It is specified in 0001 (see oid => '8810').
> > >
> > > What I meant was that I don't think these wrapper functions are
> > > needed. I think you can just do:
> > >
> > > ```
> > > +{ oid => '8811', descr => 'sort array',
> > > + proname => 'array_sort', prorettype => 'anyarray',
> > > + proargtypes => 'anyarray bool', prosrc => 'array_sort'}, <--
> > > array_sort is used directly in `prosrc`
> > > ```
> > >
> > > ... unless I'm missing something.
> >
> > There is a opr sanity check for this[1], if we remove these wrapper functions,
> > regression test will fail with:
> >
> > - oid | proname | oid | proname
> > ------+---------+-----+---------
> > -(0 rows)
> > + oid | proname | oid | proname
> > +------+------------+------+------------
> > + 8811 | array_sort | 8812 | array_sort
> > + 8810 | array_sort | 8811 | array_sort
> > + 8810 | array_sort | 8812 | array_sort
> > +(3 rows)
> >
> >
> > [1]:
> >
> > -- Considering only built-in procs (prolang = 12), look for multiple uses
> > -- of the same internal function (ie, matching prosrc fields). It's OK to
> > -- have several entries with different pronames for the same internal function,
> > -- but conflicts in the number of arguments and other critical items should
> > -- be complained of. (We don't check data types here; see next query.)
> > -- Note: ignore aggregate functions here, since they all point to the same
> > -- dummy built-in function.
> >
> > SELECT p1.oid, p1.proname, p2.oid, p2.proname
> > FROM pg_proc AS p1, pg_proc AS p2
> > WHERE p1.oid < p2.oid AND
> > p1.prosrc = p2.prosrc AND
> > p1.prolang = 12 AND p2.prolang = 12 AND
> > (p1.prokind != 'a' OR p2.prokind != 'a') AND
> > (p1.prolang != p2.prolang OR
> > p1.prokind != p2.prokind OR
> > p1.prosecdef != p2.prosecdef OR
> > p1.proleakproof != p2.proleakproof OR
> > p1.proisstrict != p2.proisstrict OR
> > p1.proretset != p2.proretset OR
> > p1.provolatile != p2.provolatile OR
> > p1.pronargs != p2.pronargs);
> >
> > >
> > > --
> > > Best regards,
> > > Aleksander Alekseev
> >
> >
> >
> > --
> > Regards
> > Junwang Zhao
>
> CFbot failed with doc build, v10 fixed that.
>
> --
> Regards
> Junwang Zhao
Rebase needed due to array_reverse committed, PFA v11.
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v11-0002-support-sort-order-and-nullsfirst-flag.patch (8.7K, ../../CAEG8a3Kah4J9q26KXY+p1UtWdm39hsZ6DsoBJbUdmB2yFXALBA@mail.gmail.com/2-v11-0002-support-sort-order-and-nullsfirst-flag.patch)
download | inline diff:
From 76ed69b5005e8445d68da4da45994748ff83e0ca Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Sat, 26 Oct 2024 03:53:17 +0000
Subject: [PATCH v11 2/2] support sort order and nullsfirst flag
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 6 +++-
src/backend/utils/adt/array_userfuncs.c | 42 ++++++++++++++++++++-----
src/include/catalog/pg_proc.dat | 6 ++++
src/test/regress/expected/arrays.out | 38 ++++++++++++++++++++++
src/test/regress/sql/arrays.sql | 10 ++++++
5 files changed, 94 insertions(+), 8 deletions(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 9e019e18cb..de66991d3d 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20443,12 +20443,16 @@ SELECT NULLIF(value, '(none)') ...
<indexterm>
<primary>array_sort</primary>
</indexterm>
- <function>array_sort</function> ( <type>anyarray</type> )
+ <function>array_sort</function> ( <type>anyarray</type> <optional>, <parameter>is_ascending</parameter> <type>boolean</type> <optional>, <parameter>nulls_first</parameter> <type>boolean</type></optional></optional>)
<returnvalue>anyarray</returnvalue>
</para>
<para>
Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ If <parameter>is_ascending</parameter> is true then sort by ascending order, otherwise descending order.
+ <parameter>is_ascending</parameter> defaults to true.
+ If <parameter>nulls_first</parameter> is true then nulls appear before non-null values, otherwise nulls appear after non-null values.
+ <parameter>nulls_first</parameter> defaults to the opposite of <parameter>is_ascending</parameter> if not provided.
</para>
<para>
<literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index a45586c892..df9822d977 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1810,7 +1810,6 @@ typedef struct ArraySortCachedInfo
* array_sort
*
* Sorts the first dimension of the array.
- * The sort order is determined by the "<" operator of the element type.
*/
Datum
array_sort(PG_FUNCTION_ARGS)
@@ -1826,10 +1825,23 @@ array_sort(PG_FUNCTION_ARGS)
Datum value;
bool isnull;
ArrayBuildStateAny *astate = NULL;
+ bool is_ascending = true;
+ bool nulls_first = false;
if (ARR_NDIM(array) < 1)
PG_RETURN_ARRAYTYPE_P(array);
+ if (PG_NARGS() > 1)
+ {
+ is_ascending = PG_GETARG_BOOL(1);
+
+ /*
+ * If nulls_first not provided, it defaults to the opposite of
+ * is_ascending.
+ */
+ nulls_first = PG_NARGS() > 2 ? PG_GETARG_BOOL(2) : !is_ascending;
+ }
+
elmtyp = ARR_ELEMTYPE(array);
cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
if (cache_info == NULL)
@@ -1847,8 +1859,10 @@ array_sort(PG_FUNCTION_ARGS)
typentry = cache_info->typentry;
if (typentry == NULL || typentry->type_id != elmtyp)
{
- typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(elmtyp,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1868,8 +1882,10 @@ array_sort(PG_FUNCTION_ARGS)
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
format_type_be(elmtyp))));
- typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(array_type,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1879,9 +1895,9 @@ array_sort(PG_FUNCTION_ARGS)
}
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
- typentry->lt_opr,
+ is_ascending ? typentry->lt_opr : typentry->gt_opr,
collation,
- false, work_mem, NULL, false);
+ nulls_first, work_mem, NULL, false);
array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
while (array_iterate(array_iterator, &value, &isnull))
@@ -1907,3 +1923,15 @@ array_sort(PG_FUNCTION_ARGS)
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
+
+Datum
+array_sort_order_nulls_first(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 55b52e138a..ad6de57e57 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1747,6 +1747,12 @@
{ oid => '8810', descr => 'sort array',
proname => 'array_sort', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort_order'},
+{ oid => '8812', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool bool', prosrc => 'array_sort_order_nulls_first'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 846d6acd85..b0cc38581e 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2765,6 +2765,44 @@ SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo,NULL}
(1 row)
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+ array_sort
+--------------------------------
+ {NULL,1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+ array_sort
+--------------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1,NULL}
+(1 row)
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
array_sort
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index b1f9c97e5c..bc046d9fe4 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -842,5 +842,15 @@ SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+
-- multidimensional array tests
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
--
2.39.5
[application/octet-stream] v11-0001-general-purpose-array_sort.patch (10.4K, ../../CAEG8a3Kah4J9q26KXY+p1UtWdm39hsZ6DsoBJbUdmB2yFXALBA@mail.gmail.com/3-v11-0001-general-purpose-array_sort.patch)
download | inline diff:
From 981ae3c7e4d3899c00b61b02b98559549ce3bc84 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v11 1/2] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 110 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 3 +
src/test/regress/expected/arrays.out | 38 ++++++
.../regress/expected/collate.icu.utf8.out | 13 +++
src/test/regress/sql/arrays.sql | 12 +-
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 198 insertions(+), 1 deletion(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 223d869f8c..9e019e18cb 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20438,6 +20438,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> )
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
+ The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index d053808f6e..a45586c892 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1797,3 +1799,111 @@ array_reverse(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry;
+ TypeCacheEntry *array_typentry;
+} ArraySortCachedInfo;
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ARR_NDIM(array) == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index a38e20f5d9..55b52e138a 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,9 @@
{ oid => '8686', descr => 'reverse array',
proname => 'array_reverse', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_reverse' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 0b61fb5bb7..846d6acd85 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2734,3 +2734,41 @@ SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
{{7,8},{5,6},{3,4},{1,2}}
(1 row)
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 691cff4a12..b1f9c97e5c 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -833,4 +833,14 @@ SELECT array_reverse('{}'::int[]);
SELECT array_reverse('{1}'::int[]);
SELECT array_reverse('{1,2}'::int[]);
SELECT array_reverse('{1,2,3,NULL,4,5,6}'::int[]);
-SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
\ No newline at end of file
+SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 171a7dd5d2..eceba9605e 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-04 05:46 Michael Paquier <[email protected]>
parent: Junwang Zhao <[email protected]>
1 sibling, 2 replies; 166+ messages in thread
From: Michael Paquier @ 2024-11-04 05:46 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Sun, Nov 03, 2024 at 11:33:05AM +0800, Junwang Zhao wrote:
> Rebase needed due to array_reverse committed, PFA v11.
There has been another conflict since you have posted this version
(noticed that after my business in 027124a872d7). I have looked at
0001.
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
There is no point in doing a sort if the array has only one element.
You can add a check based on "ARR_DIMS(array)[0] < 2" to achieve that.
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
The patch introduces two error paths based on the fact that ordering
operators could not be found depending on a data type that lacks the
ordering operator and the array ordering operator part. It is right
to issue an error if these are lacking, like the various stats paths.
Should we have some regression tests with specific data types for
these errors, though? The stats paths don't care much about these
error cases, but it does not mean that we should not care about them.
In short, let's have negative test coverage if we can.
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry;
+ TypeCacheEntry *array_typentry;
+} ArraySortCachedInfo;
Let's put that at the top of the file, with a comment about how it
links to array_sort() for the caching with fn_extra. Let's also
document the meaning of the fields.
FWIW, I am confused by this implementation, where you have to allocate
the two TypeCacheEntry because of the fact that you have to deal with
the 1-dimension case and the multi-dimension case. In the context of
a single function call, why do you need both typentry and
array_typentry, actually? Wouldn't it be enough to use one typentry
that points to the typcache, meaning that you don't really need to use
the extra business with fn_mcxt, no? If you require both (because I
may be wrong), perhaps you should have a regression test that's able
to break when removing array_typentry, changing the code to only rely
on typentry. Note: I have just removed array_typentry in a quick
test, current coverage was happy about it. Feel free to prove me
wrong.
Agreed that the function should be immutable. The results are fixed
depending on the input even with the COLLATE clauses appended.
Let's add something when there is only one element in the first
dimension of the array, say two cases one with an int and one with an
array of ints like:
SELECT array_sort('{1}'::int[]);
SELECT array_sort('{{1}}'::int[]);
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-04 07:16 jian he <[email protected]>
parent: Michael Paquier <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: jian he @ 2024-11-04 07:16 UTC (permalink / raw)
To: Michael Paquier <[email protected]>; +Cc: Junwang Zhao <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Mon, Nov 4, 2024 at 1:46 PM Michael Paquier <[email protected]> wrote:
>
>
> + typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
> + if (!OidIsValid(typentry->lt_opr))
> + ereport(ERROR,
> + (errcode(ERRCODE_UNDEFINED_FUNCTION),
> + errmsg("could not identify ordering operator for type %s",
> + format_type_be(elmtyp))));
>
> The patch introduces two error paths based on the fact that ordering
> operators could not be found depending on a data type that lacks the
> ordering operator and the array ordering operator part. It is right
> to issue an error if these are lacking, like the various stats paths.
> Should we have some regression tests with specific data types for
> these errors, though? The stats paths don't care much about these
> error cases, but it does not mean that we should not care about them.
> In short, let's have negative test coverage if we can.
>
select distinct oprleft::regtype from pg_operator where oprname = '='
and oprleft = oprright
except all
select distinct oprleft::regtype from pg_operator where oprname = '<'
and oprleft = oprright;
returns
hstore
cid
aclitem
xid
line
simple tests case using xid data type would be
SELECT array_sort('{{1,2,3}}'::xid[]);
> +typedef struct ArraySortCachedInfo
> +{
> + TypeCacheEntry *typentry;
> + TypeCacheEntry *array_typentry;
> +} ArraySortCachedInfo;
>
> Let's put that at the top of the file, with a comment about how it
> links to array_sort() for the caching with fn_extra. Let's also
> document the meaning of the fields.
>
> FWIW, I am confused by this implementation, where you have to allocate
> the two TypeCacheEntry because of the fact that you have to deal with
> the 1-dimension case and the multi-dimension case. In the context of
> a single function call, why do you need both typentry and
> array_typentry, actually? Wouldn't it be enough to use one typentry
> that points to the typcache, meaning that you don't really need to use
> the extra business with fn_mcxt, no? If you require both (because I
> may be wrong), perhaps you should have a regression test that's able
> to break when removing array_typentry, changing the code to only rely
> on typentry. Note: I have just removed array_typentry in a quick
> test, current coverage was happy about it. Feel free to prove me
> wrong.
>
drop table if exists t;
CREATE TABLE t (a int[]);
insert into t values ('{1,3}'),('{1,2,3}'),('{11}');
insert into t values ('{{1,12}}'), ('{{4,3}}');
SELECT array_sort(a) from t;
In the above case,
tuplesort_begin_datum needs the int type information and int[] type information.
otherwise the cached TypeCacheEntry is being used to sort mult-dimension array,
which will make the result false.
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-04 11:33 Dean Rasheed <[email protected]>
parent: Junwang Zhao <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: Dean Rasheed @ 2024-11-04 11:33 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Sun, 3 Nov 2024 at 03:33, Junwang Zhao <[email protected]> wrote:
>
> PFA v11.
>
Testing this with an array with non-default lower bounds, it fails to
preserve the array bounds, which I think it should (note:
array_reverse() and array_shuffle() do preserve the bounds):
SELECT array_reverse(a), array_shuffle(a), array_sort(a)
FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
-[ RECORD 1 ]-+-------------------------------------
array_reverse | [10:12][20:21]={{3,4},{10,20},{1,2}}
array_shuffle | [10:12][20:21]={{10,20},{3,4},{1,2}}
array_sort | [1:3][20:21]={{1,2},{3,4},{10,20}}
Regards,
Dean
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-05 01:12 Michael Paquier <[email protected]>
parent: jian he <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Michael Paquier @ 2024-11-05 01:12 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Junwang Zhao <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Mon, Nov 04, 2024 at 03:16:35PM +0800, jian he wrote:
> drop table if exists t;
> CREATE TABLE t (a int[]);
> insert into t values ('{1,3}'),('{1,2,3}'),('{11}');
> insert into t values ('{{1,12}}'), ('{{4,3}}');
> SELECT array_sort(a) from t;
>
> In the above case,
> tuplesort_begin_datum needs the int type information and int[] type information.
> otherwise the cached TypeCacheEntry is being used to sort mult-dimension array,
> which will make the result false.
All these behaviors need more extensive testing.
This brings me an extra question around the caching. Would the
sorting be able to behave correctly when feeding to a single
array_sort() context array values that have multiple COLLATE clauses?
Or merge_collation_state() would be smart enough to make sure that
collation conflicts never happen to begin with? I am wondering if we
should worry about multiple VALUES, CTEs, or PL functions where
array_sort() could be fed into its cache values that lead to
unpredictible results for some values. This stuff should perhaps have
more testing around such behaviors, stressing what kind of
interactions we have between the sorting of multiple values and the
caching, in the context of a single array_sort() call.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-05 07:12 jian he <[email protected]>
parent: Dean Rasheed <[email protected]>
0 siblings, 2 replies; 166+ messages in thread
From: jian he @ 2024-11-05 07:12 UTC (permalink / raw)
To: Dean Rasheed <[email protected]>; +Cc: Junwang Zhao <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Mon, Nov 4, 2024 at 7:34 PM Dean Rasheed <[email protected]> wrote:
>
> Testing this with an array with non-default lower bounds, it fails to
> preserve the array bounds, which I think it should (note:
> array_reverse() and array_shuffle() do preserve the bounds):
>
> SELECT array_reverse(a), array_shuffle(a), array_sort(a)
> FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
>
> -[ RECORD 1 ]-+-------------------------------------
> array_reverse | [10:12][20:21]={{3,4},{10,20},{1,2}}
> array_shuffle | [10:12][20:21]={{10,20},{3,4},{1,2}}
> array_sort | [1:3][20:21]={{1,2},{3,4},{10,20}}
>
if i understand it correctly,
array_create_iterator cannot cope with top dimension bound information.
since input array arguments already have dims, lbs information.
so at the end of array_sort directly copy
from the input array argument to astate.
tuplesort_performsort won't need array bounds, we should be safe?
v12-0001 same as v11-0001-general-purpose-array_sort.patch, only
resolve git conflict
v12-0002 preserve array bound information.
v12-0003 cache ArrayMetaState.
after v12-0003 now
typedef struct ArraySortCachedInfo
{
TypeCacheEntry *typentry;
TypeCacheEntry *array_typentry;
ArrayMetaState array_meta;
} ArraySortCachedInfo;
function array_create_iterator, get_typlenbyvalalign
will do cache search, we can cache ArrayMetaState.
so multiple array_create_iterator calls won't need to call get_typlenbyvalalign.
every time.
0002, I also have a 3 dimensional array test.
create table t(a int[]);
insert into t values ('[-1:-0]={7,1}'::int[]),
('[-2:-0][20:21]={{1,2},{10,20},{1,-4}}'),
('[-2:-0][20:22]={{-11,2,-1},{-11,2, 1},{-11,-4, 10}}'),
('[-13:-10][0:1][20:22]={
{{1,2,112},{1,2,-123}},
{{10,-20,1},{11,123,3}},
{{10,-20,1},{11,-123,-9}},
{{1,2,-11},{1,2,211}}}'::int[]);
SELECT array_sort(t.a) from t;
SELECT array_sort((t.a) [-13:-10][0:1][21:22]) from t where array_ndims(a) = 3;
SELECT array_sort((t.a) [-13:-11][0:1][21:22]) from t where array_ndims(a) = 3;
SELECT array_sort((t.a) [-13:-11][0:0][20:21]) from t where array_ndims(a) = 3;
The test output is ok to me.
Attachments:
[application/x-patch] v12-0003-array_sort-cache-ArrayMetaState.patch (2.2K, ../../CACJufxGMCALTkLS0MR9SYHmMb3Y0xwGS8nvSQmSk6zTerO=aQA@mail.gmail.com/2-v12-0003-array_sort-cache-ArrayMetaState.patch)
download | inline diff:
From caa24ca549c01e99b3873860dcffe0d8161345d4 Mon Sep 17 00:00:00 2001
From: jian he <[email protected]>
Date: Tue, 5 Nov 2024 14:34:02 +0800
Subject: [PATCH v12 3/3] array_sort cache ArrayMetaState
---
src/backend/utils/adt/array_userfuncs.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 436f761614..c14ef1ed90 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1804,6 +1804,7 @@ typedef struct ArraySortCachedInfo
{
TypeCacheEntry *typentry;
TypeCacheEntry *array_typentry;
+ ArrayMetaState array_meta;
} ArraySortCachedInfo;
/*
@@ -1848,6 +1849,7 @@ array_sort(PG_FUNCTION_ARGS)
sizeof(ArraySortCachedInfo));
cache_info->typentry = NULL;
cache_info->array_typentry = NULL;
+ memset(&cache_info->array_meta, 0, sizeof(ArrayMetaState));
fcinfo->flinfo->fn_extra = (void *) cache_info;
}
@@ -1864,6 +1866,10 @@ array_sort(PG_FUNCTION_ARGS)
errmsg("could not identify ordering operator for type %s",
format_type_be(elmtyp))));
cache_info->typentry = typentry;
+ cache_info->array_meta.element_type = elmtyp;
+ cache_info->array_meta.typlen = typentry->typlen;
+ cache_info->array_meta.typbyval = typentry->typbyval;
+ cache_info->array_meta.typalign = typentry->typalign;
}
}
else
@@ -1886,6 +1892,11 @@ array_sort(PG_FUNCTION_ARGS)
format_type_be(array_type))));
cache_info->array_typentry = typentry;
}
+ cache_info->array_meta.element_type = elmtyp;
+ get_typlenbyvalalign(elmtyp,
+ &cache_info->array_meta.typlen,
+ &cache_info->array_meta.typbyval,
+ &cache_info->array_meta.typalign);
}
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
@@ -1893,7 +1904,7 @@ array_sort(PG_FUNCTION_ARGS)
collation,
false, work_mem, NULL, false);
- array_iterator = array_create_iterator(array, ndim - 1, NULL);
+ array_iterator = array_create_iterator(array, ndim - 1, &cache_info->array_meta);
while (array_iterate(array_iterator, &value, &isnull))
{
tuplesort_putdatum(tuplesortstate, value, isnull);
--
2.34.1
[application/x-patch] v12-0002-array_sort-preserve-array-dimenion-and-bound-inf.patch (7.4K, ../../CACJufxGMCALTkLS0MR9SYHmMb3Y0xwGS8nvSQmSk6zTerO=aQA@mail.gmail.com/3-v12-0002-array_sort-preserve-array-dimenion-and-bound-inf.patch)
download | inline diff:
From d7c2956893222bb025822810860a01539ca3ec85 Mon Sep 17 00:00:00 2001
From: jian he <[email protected]>
Date: Tue, 5 Nov 2024 14:53:42 +0800
Subject: [PATCH v12 2/3] array_sort preserve array dimenion and bound info
---
src/backend/utils/adt/array_userfuncs.c | 30 +++++++++++++--
src/backend/utils/adt/arrayfuncs.c | 3 +-
src/include/utils/array.h | 1 +
src/test/regress/expected/arrays.out | 49 +++++++++++++++++++++++++
src/test/regress/sql/arrays.sql | 18 ++++++++-
5 files changed, 96 insertions(+), 5 deletions(-)
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index a45586c892..436f761614 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1826,8 +1826,18 @@ array_sort(PG_FUNCTION_ARGS)
Datum value;
bool isnull;
ArrayBuildStateAny *astate = NULL;
+ int ndim,
+ *dims,
+ *lbs;
- if (ARR_NDIM(array) < 1)
+ ndim = ARR_NDIM(array);
+ dims = ARR_DIMS(array);
+ lbs = ARR_LBOUND(array);
+
+ if (ndim < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ if(dims[0] < 2)
PG_RETURN_ARRAYTYPE_P(array);
elmtyp = ARR_ELEMTYPE(array);
@@ -1841,7 +1851,7 @@ array_sort(PG_FUNCTION_ARGS)
fcinfo->flinfo->fn_extra = (void *) cache_info;
}
- if (ARR_NDIM(array) == 1)
+ if (ndim == 1)
{
/* Finds the ordering operator for the type for 1-D arrays */
typentry = cache_info->typentry;
@@ -1883,7 +1893,7 @@ array_sort(PG_FUNCTION_ARGS)
collation,
false, work_mem, NULL, false);
- array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ array_iterator = array_create_iterator(array, ndim - 1, NULL);
while (array_iterate(array_iterator, &value, &isnull))
{
tuplesort_putdatum(tuplesortstate, value, isnull);
@@ -1903,6 +1913,20 @@ array_sort(PG_FUNCTION_ARGS)
tuplesort_end(tuplesortstate);
+ if (astate->arraystate != NULL)
+ {
+ memcpy(astate->arraystate->dims, dims, ndim * sizeof(int));
+ memcpy(astate->arraystate->lbs, lbs, ndim * sizeof(int));
+ Assert(ndim == astate->arraystate->ndims);
+ }
+ else
+ {
+ astate->scalarstate->lbs = lbs[0];
+
+ /* zero dimension won't reach here */
+ Assert(ndim == 1);
+ }
+
/* Avoid leaking memory when handed toasted input */
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
diff --git a/src/backend/utils/adt/arrayfuncs.c b/src/backend/utils/adt/arrayfuncs.c
index a715e7e0b8..ffdf3449e9 100644
--- a/src/backend/utils/adt/arrayfuncs.c
+++ b/src/backend/utils/adt/arrayfuncs.c
@@ -5335,6 +5335,7 @@ initArrayResultWithSize(Oid element_type, MemoryContext rcontext,
&astate->typbyval,
&astate->typalign);
+ astate->lbs = 1; /* default lower-bound value set to 1, see array_in also */
return astate;
}
@@ -5867,7 +5868,7 @@ makeArrayResultAny(ArrayBuildStateAny *astate,
/* If no elements were presented, we want to create an empty array */
ndims = (astate->scalarstate->nelems > 0) ? 1 : 0;
dims[0] = astate->scalarstate->nelems;
- lbs[0] = 1;
+ lbs[0] = astate->scalarstate->lbs;
result = makeMdArrayResult(astate->scalarstate, ndims, dims, lbs,
rcontext, release);
diff --git a/src/include/utils/array.h b/src/include/utils/array.h
index 157cc0e4c6..7b7937d494 100644
--- a/src/include/utils/array.h
+++ b/src/include/utils/array.h
@@ -193,6 +193,7 @@ typedef struct ArrayBuildState
int nelems; /* number of valid entries in above arrays */
Oid element_type; /* data type of the Datums */
int16 typlen; /* needed info about datatype */
+ int lbs; /* one dimension, one lower-bound is OK*/
bool typbyval;
char typalign;
bool private_cxt; /* use private memory context */
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 846d6acd85..9e371902d3 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2735,6 +2735,18 @@ SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
(1 row)
-- array_sort
+SELECT array_sort('{1}'::int[]);
+ array_sort
+------------
+ {1}
+(1 row)
+
+SELECT array_sort('{{2,1}}'::int[]);
+ array_sort
+------------
+ {{2,1}}
+(1 row)
+
SELECT array_sort('{}'::int[]);
array_sort
------------
@@ -2772,3 +2784,40 @@ SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
{{2,1},{2,4},{6,5}}
(1 row)
+create table t(a int[]);
+insert into t values ('[-1:-0]={7,1}'::int[]),
+('[-2:-0][20:21]={{1,2},{10,20},{1,-4}}'),
+('[-2:-0][20:22]={{-11,2,-1},{-11,2, 1},{-11,-4, 10}}'),
+('[-13:-10][0:1][20:22]={
+{{1,2,112},{1,2,-123}},
+{{10,-20,1},{11,123,3}},
+{{10,-20,1},{11,-123,-9}},
+{{1,2,-11},{1,2,211}}}'::int[]);
+SELECT array_sort(t.a) from t;
+ array_sort
+------------------------------------------------------------------------------------------------------------------------
+ [-1:0]={1,7}
+ [-2:0][20:21]={{1,-4},{1,2},{10,20}}
+ [-2:0][20:22]={{-11,-4,10},{-11,2,-1},{-11,2,1}}
+ [-13:-10][0:1][20:22]={{{1,2,-11},{1,2,211}},{{1,2,112},{1,2,-123}},{{10,-20,1},{11,-123,-9}},{{10,-20,1},{11,123,3}}}
+(4 rows)
+
+SELECT array_sort((t.a) [-13:-10][0:1][21:22]) from t where array_ndims(a) = 3;
+ array_sort
+------------------------------------------------------------------------------
+ {{{-20,1},{-123,-9}},{{-20,1},{123,3}},{{2,-11},{2,211}},{{2,112},{2,-123}}}
+(1 row)
+
+SELECT array_sort((t.a) [-13:-11][0:1][21:22]) from t where array_ndims(a) = 3;
+ array_sort
+------------------------------------------------------------
+ {{{-20,1},{-123,-9}},{{-20,1},{123,3}},{{2,112},{2,-123}}}
+(1 row)
+
+SELECT array_sort((t.a) [-13:-11][0:0][20:21]) from t where array_ndims(a) = 3;
+ array_sort
+---------------------------------
+ {{{1,2}},{{10,-20}},{{10,-20}}}
+(1 row)
+
+drop table t;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 17205f483b..2a40c401b5 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -836,6 +836,8 @@ SELECT array_reverse('{1,2,3,NULL,4,5,6}'::int[]);
SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
-- array_sort
+SELECT array_sort('{1}'::int[]);
+SELECT array_sort('{{2,1}}'::int[]);
SELECT array_sort('{}'::int[]);
SELECT array_sort('{1,3,5,2,4,6}'::int[]);
SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
@@ -843,4 +845,18 @@ SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
-- multidimensional array tests
-SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
\ No newline at end of file
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+create table t(a int[]);
+insert into t values ('[-1:-0]={7,1}'::int[]),
+('[-2:-0][20:21]={{1,2},{10,20},{1,-4}}'),
+('[-2:-0][20:22]={{-11,2,-1},{-11,2, 1},{-11,-4, 10}}'),
+('[-13:-10][0:1][20:22]={
+{{1,2,112},{1,2,-123}},
+{{10,-20,1},{11,123,3}},
+{{10,-20,1},{11,-123,-9}},
+{{1,2,-11},{1,2,211}}}'::int[]);
+SELECT array_sort(t.a) from t;
+SELECT array_sort((t.a) [-13:-10][0:1][21:22]) from t where array_ndims(a) = 3;
+SELECT array_sort((t.a) [-13:-11][0:1][21:22]) from t where array_ndims(a) = 3;
+SELECT array_sort((t.a) [-13:-11][0:0][20:21]) from t where array_ndims(a) = 3;
+drop table t;
\ No newline at end of file
--
2.34.1
[application/x-patch] v12-0001-general-purpose-array_sort.patch (10.3K, ../../CACJufxGMCALTkLS0MR9SYHmMb3Y0xwGS8nvSQmSk6zTerO=aQA@mail.gmail.com/4-v12-0001-general-purpose-array_sort.patch)
download | inline diff:
From ea85acab5039ef8d23e1e994c9d1d4ae79bdb29e Mon Sep 17 00:00:00 2001
From: jian he <[email protected]>
Date: Tue, 5 Nov 2024 14:40:56 +0800
Subject: [PATCH v12 1/3] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 110 ++++++++++++++++++
src/include/catalog/pg_proc.dat | 3 +
src/test/regress/expected/arrays.out | 38 ++++++
.../regress/expected/collate.icu.utf8.out | 13 +++
src/test/regress/sql/arrays.sql | 10 ++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
8 files changed, 197 insertions(+)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 73979f20ff..748611dee5 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20438,6 +20438,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> )
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
+ The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index d053808f6e..a45586c892 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -1797,3 +1799,111 @@ array_reverse(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry;
+ TypeCacheEntry *array_typentry;
+} ArraySortCachedInfo;
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+
+ if (ARR_NDIM(array) < 1)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ARR_NDIM(array) == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ARR_NDIM(array) - 1, NULL);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index f23321a41f..54ebbbd135 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,9 @@
{ oid => '8686', descr => 'reverse array',
proname => 'array_reverse', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_reverse' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 0b61fb5bb7..846d6acd85 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2734,3 +2734,41 @@ SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
{{7,8},{5,6},{3,4},{1,2}}
(1 row)
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 03cc8cfcd9..17205f483b 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -834,3 +834,13 @@ SELECT array_reverse('{1}'::int[]);
SELECT array_reverse('{1,2}'::int[]);
SELECT array_reverse('{1,2,3,NULL,4,5,6}'::int[]);
SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
\ No newline at end of file
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1847bbfa95..e740e449f5 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.34.1
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-05 12:30 Junwang Zhao <[email protected]>
parent: jian he <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-11-05 12:30 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Dean Rasheed <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi jian,
On Tue, Nov 5, 2024 at 3:13 PM jian he <[email protected]> wrote:
>
> On Mon, Nov 4, 2024 at 7:34 PM Dean Rasheed <[email protected]> wrote:
> >
> > Testing this with an array with non-default lower bounds, it fails to
> > preserve the array bounds, which I think it should (note:
> > array_reverse() and array_shuffle() do preserve the bounds):
> >
> > SELECT array_reverse(a), array_shuffle(a), array_sort(a)
> > FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
> >
> > -[ RECORD 1 ]-+-------------------------------------
> > array_reverse | [10:12][20:21]={{3,4},{10,20},{1,2}}
> > array_shuffle | [10:12][20:21]={{10,20},{3,4},{1,2}}
> > array_sort | [1:3][20:21]={{1,2},{3,4},{10,20}}
> >
>
> if i understand it correctly,
> array_create_iterator cannot cope with top dimension bound information.
> since input array arguments already have dims, lbs information.
> so at the end of array_sort directly copy
> from the input array argument to astate.
>
> tuplesort_performsort won't need array bounds, we should be safe?
>
>
>
> v12-0001 same as v11-0001-general-purpose-array_sort.patch, only
> resolve git conflict
> v12-0002 preserve array bound information.
> v12-0003 cache ArrayMetaState.
>
> after v12-0003 now
> typedef struct ArraySortCachedInfo
> {
> TypeCacheEntry *typentry;
> TypeCacheEntry *array_typentry;
> ArrayMetaState array_meta;
> } ArraySortCachedInfo;
>
> function array_create_iterator, get_typlenbyvalalign
> will do cache search, we can cache ArrayMetaState.
> so multiple array_create_iterator calls won't need to call get_typlenbyvalalign.
> every time.
>
>
> 0002, I also have a 3 dimensional array test.
> create table t(a int[]);
> insert into t values ('[-1:-0]={7,1}'::int[]),
> ('[-2:-0][20:21]={{1,2},{10,20},{1,-4}}'),
> ('[-2:-0][20:22]={{-11,2,-1},{-11,2, 1},{-11,-4, 10}}'),
> ('[-13:-10][0:1][20:22]={
> {{1,2,112},{1,2,-123}},
> {{10,-20,1},{11,123,3}},
> {{10,-20,1},{11,-123,-9}},
> {{1,2,-11},{1,2,211}}}'::int[]);
> SELECT array_sort(t.a) from t;
> SELECT array_sort((t.a) [-13:-10][0:1][21:22]) from t where array_ndims(a) = 3;
> SELECT array_sort((t.a) [-13:-11][0:1][21:22]) from t where array_ndims(a) = 3;
> SELECT array_sort((t.a) [-13:-11][0:0][20:21]) from t where array_ndims(a) = 3;
>
> The test output is ok to me.
Thanks for the bounds preserve solution, I just looked at 0002,
+ if (astate->arraystate != NULL)
+ {
+ memcpy(astate->arraystate->dims, dims, ndim * sizeof(int));
+ memcpy(astate->arraystate->lbs, lbs, ndim * sizeof(int));
+ Assert(ndim == astate->arraystate->ndims);
+ }
It seems to me we only need to set astate->arraystate->lbs[0] = lbs[0] ?
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-06 07:10 jian he <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: jian he @ 2024-11-06 07:10 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: Dean Rasheed <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Tue, Nov 5, 2024 at 8:30 PM Junwang Zhao <[email protected]> wrote:
>
>
> Thanks for the bounds preserve solution, I just looked at 0002,
>
> + if (astate->arraystate != NULL)
> + {
> + memcpy(astate->arraystate->dims, dims, ndim * sizeof(int));
> + memcpy(astate->arraystate->lbs, lbs, ndim * sizeof(int));
> + Assert(ndim == astate->arraystate->ndims);
> + }
>
> It seems to me we only need to set astate->arraystate->lbs[0] = lbs[0] ?
>
yes.
> + memcpy(astate->arraystate->dims, dims, ndim * sizeof(int));
thinking about it, this is wrong. we should just do Assert
for(int i = 0; i < ndim; i++)
{
Assert(astate->arraystate->dims[i] == dims[i]);
}
or just remove
memcpy(astate->arraystate->dims, dims, ndim * sizeof(int));
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-07 13:56 Junwang Zhao <[email protected]>
parent: Michael Paquier <[email protected]>
1 sibling, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-11-07 13:56 UTC (permalink / raw)
To: Michael Paquier <[email protected]>; +Cc: Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Michael,
On Mon, Nov 4, 2024 at 1:46 PM Michael Paquier <[email protected]> wrote:
>
> On Sun, Nov 03, 2024 at 11:33:05AM +0800, Junwang Zhao wrote:
> > Rebase needed due to array_reverse committed, PFA v11.
>
> There has been another conflict since you have posted this version
> (noticed that after my business in 027124a872d7). I have looked at
> 0001.
>
> + if (ARR_NDIM(array) < 1)
> + PG_RETURN_ARRAYTYPE_P(array);
> There is no point in doing a sort if the array has only one element.
> You can add a check based on "ARR_DIMS(array)[0] < 2" to achieve that.
Yeah, this is reasonable but one case I can't be sure:
SELECT array_sort('{{2,3,4}}'::xid[]);
This will return the array as is, but xid doesn't have a LT_OPR, should
I error out in this case? like:
could not identify ordering operator for type xid[]
>
> + typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
> + if (!OidIsValid(typentry->lt_opr))
> + ereport(ERROR,
> + (errcode(ERRCODE_UNDEFINED_FUNCTION),
> + errmsg("could not identify ordering operator for type %s",
> + format_type_be(elmtyp))));
>
> The patch introduces two error paths based on the fact that ordering
> operators could not be found depending on a data type that lacks the
> ordering operator and the array ordering operator part. It is right
> to issue an error if these are lacking, like the various stats paths.
> Should we have some regression tests with specific data types for
> these errors, though? The stats paths don't care much about these
> error cases, but it does not mean that we should not care about them.
> In short, let's have negative test coverage if we can.
>
> +typedef struct ArraySortCachedInfo
> +{
> + TypeCacheEntry *typentry;
> + TypeCacheEntry *array_typentry;
> +} ArraySortCachedInfo;
>
> Let's put that at the top of the file, with a comment about how it
> links to array_sort() for the caching with fn_extra. Let's also
> document the meaning of the fields.
Will fix it in the following patch set.
>
> FWIW, I am confused by this implementation, where you have to allocate
> the two TypeCacheEntry because of the fact that you have to deal with
> the 1-dimension case and the multi-dimension case. In the context of
> a single function call, why do you need both typentry and
> array_typentry, actually? Wouldn't it be enough to use one typentry
> that points to the typcache, meaning that you don't really need to use
> the extra business with fn_mcxt, no? If you require both (because I
> may be wrong), perhaps you should have a regression test that's able
> to break when removing array_typentry, changing the code to only rely
> on typentry. Note: I have just removed array_typentry in a quick
> test, current coverage was happy about it. Feel free to prove me
> wrong.
>
> Agreed that the function should be immutable. The results are fixed
> depending on the input even with the COLLATE clauses appended.
>
> Let's add something when there is only one element in the first
> dimension of the array, say two cases one with an int and one with an
> array of ints like:
> SELECT array_sort('{1}'::int[]);
> SELECT array_sort('{{1}}'::int[]);
Will add.
> --
> Michael
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-07 14:06 Junwang Zhao <[email protected]>
parent: Michael Paquier <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Junwang Zhao @ 2024-11-07 14:06 UTC (permalink / raw)
To: Michael Paquier <[email protected]>; +Cc: jian he <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Tue, Nov 5, 2024 at 9:13 AM Michael Paquier <[email protected]> wrote:
>
> On Mon, Nov 04, 2024 at 03:16:35PM +0800, jian he wrote:
> > drop table if exists t;
> > CREATE TABLE t (a int[]);
> > insert into t values ('{1,3}'),('{1,2,3}'),('{11}');
> > insert into t values ('{{1,12}}'), ('{{4,3}}');
> > SELECT array_sort(a) from t;
> >
> > In the above case,
> > tuplesort_begin_datum needs the int type information and int[] type information.
> > otherwise the cached TypeCacheEntry is being used to sort mult-dimension array,
> > which will make the result false.
>
> All these behaviors need more extensive testing.
>
> This brings me an extra question around the caching. Would the
> sorting be able to behave correctly when feeding to a single
> array_sort() context array values that have multiple COLLATE clauses?
> Or merge_collation_state() would be smart enough to make sure that
> collation conflicts never happen to begin with? I am wondering if we
> should worry about multiple VALUES, CTEs, or PL functions where
> array_sort() could be fed into its cache values that lead to
> unpredictible results for some values. This stuff should perhaps have
> more testing around such behaviors, stressing what kind of
> interactions we have between the sorting of multiple values and the
> caching, in the context of a single array_sort() call.
I'm afraid this can not be achieved in my current implementation, a simple
case is:
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[]);
{Abc,bar,bbc,CCC,foo,NULL}
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo,NULL}
SELECT array_sort(a COLLATE "C") FROM (VALUES
('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C"),
('{foo,bar,null,CCC,Abc,bbc}'::text[])) v(a);
{Abc,CCC,bar,bbc,foo,NULL}
{Abc,CCC,bar,bbc,foo,NULL}
Maybe add some documents to specify this?
> --
> Michael
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-07 14:26 Junwang Zhao <[email protected]>
parent: jian he <[email protected]>
1 sibling, 0 replies; 166+ messages in thread
From: Junwang Zhao @ 2024-11-07 14:26 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Dean Rasheed <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi jian,
On Tue, Nov 5, 2024 at 3:13 PM jian he <[email protected]> wrote:
>
> On Mon, Nov 4, 2024 at 7:34 PM Dean Rasheed <[email protected]> wrote:
> >
> > Testing this with an array with non-default lower bounds, it fails to
> > preserve the array bounds, which I think it should (note:
> > array_reverse() and array_shuffle() do preserve the bounds):
> >
> > SELECT array_reverse(a), array_shuffle(a), array_sort(a)
> > FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
> >
> > -[ RECORD 1 ]-+-------------------------------------
> > array_reverse | [10:12][20:21]={{3,4},{10,20},{1,2}}
> > array_shuffle | [10:12][20:21]={{10,20},{3,4},{1,2}}
> > array_sort | [1:3][20:21]={{1,2},{3,4},{10,20}}
> >
>
> if i understand it correctly,
> array_create_iterator cannot cope with top dimension bound information.
> since input array arguments already have dims, lbs information.
> so at the end of array_sort directly copy
> from the input array argument to astate.
>
> tuplesort_performsort won't need array bounds, we should be safe?
>
>
>
> v12-0001 same as v11-0001-general-purpose-array_sort.patch, only
> resolve git conflict
> v12-0002 preserve array bound information.
> v12-0003 cache ArrayMetaState.
>
> after v12-0003 now
> typedef struct ArraySortCachedInfo
> {
> TypeCacheEntry *typentry;
> TypeCacheEntry *array_typentry;
> ArrayMetaState array_meta;
> } ArraySortCachedInfo;
>
> function array_create_iterator, get_typlenbyvalalign
> will do cache search, we can cache ArrayMetaState.
> so multiple array_create_iterator calls won't need to call get_typlenbyvalalign.
> every time.
>
>
> 0002, I also have a 3 dimensional array test.
> create table t(a int[]);
> insert into t values ('[-1:-0]={7,1}'::int[]),
> ('[-2:-0][20:21]={{1,2},{10,20},{1,-4}}'),
> ('[-2:-0][20:22]={{-11,2,-1},{-11,2, 1},{-11,-4, 10}}'),
> ('[-13:-10][0:1][20:22]={
> {{1,2,112},{1,2,-123}},
> {{10,-20,1},{11,123,3}},
> {{10,-20,1},{11,-123,-9}},
> {{1,2,-11},{1,2,211}}}'::int[]);
> SELECT array_sort(t.a) from t;
> SELECT array_sort((t.a) [-13:-10][0:1][21:22]) from t where array_ndims(a) = 3;
> SELECT array_sort((t.a) [-13:-11][0:1][21:22]) from t where array_ndims(a) = 3;
> SELECT array_sort((t.a) [-13:-11][0:0][20:21]) from t where array_ndims(a) = 3;
>
> The test output is ok to me.
Merged into V13 with some change and added you as a co-author.
There is one issue left as Micheal pointed out that feeding to a single
array_sort() context array values that have multiple COLLATE clauses,
I have no idea how to resolve it ;(
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v13-0002-support-sort-order-and-nullsfirst-flag.patch (8.8K, ../../CAEG8a3KvVjgx5FRWjMq3Jai2LL+xSwDPXsYrMJuL2Jnkr9=zdA@mail.gmail.com/2-v13-0002-support-sort-order-and-nullsfirst-flag.patch)
download | inline diff:
From d74ba41d298eaf7f790a1ed4c4e9702ebb519aeb Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Sat, 26 Oct 2024 03:53:17 +0000
Subject: [PATCH v13 2/2] support sort order and nullsfirst flag
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 6 +++-
src/backend/utils/adt/array_userfuncs.c | 42 ++++++++++++++++++++-----
src/include/catalog/pg_proc.dat | 6 ++++
src/test/regress/expected/arrays.out | 38 ++++++++++++++++++++++
src/test/regress/sql/arrays.sql | 10 ++++++
5 files changed, 94 insertions(+), 8 deletions(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 748611dee5..afbda78688 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20443,12 +20443,16 @@ SELECT NULLIF(value, '(none)') ...
<indexterm>
<primary>array_sort</primary>
</indexterm>
- <function>array_sort</function> ( <type>anyarray</type> )
+ <function>array_sort</function> ( <type>anyarray</type> <optional>, <parameter>is_ascending</parameter> <type>boolean</type> <optional>, <parameter>nulls_first</parameter> <type>boolean</type></optional></optional>)
<returnvalue>anyarray</returnvalue>
</para>
<para>
Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ If <parameter>is_ascending</parameter> is true then sort by ascending order, otherwise descending order.
+ <parameter>is_ascending</parameter> defaults to true.
+ If <parameter>nulls_first</parameter> is true then nulls appear before non-null values, otherwise nulls appear after non-null values.
+ <parameter>nulls_first</parameter> defaults to the opposite of <parameter>is_ascending</parameter> if not provided.
</para>
<para>
<literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 3a50e5ca97..6e5125b46e 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1816,7 +1816,6 @@ array_reverse(PG_FUNCTION_ARGS)
* array_sort
*
* Sorts the first dimension of the array.
- * The sort order is determined by the "<" operator of the element type.
*/
Datum
array_sort(PG_FUNCTION_ARGS)
@@ -1835,6 +1834,8 @@ array_sort(PG_FUNCTION_ARGS)
int ndim,
*dims,
*lbs;
+ bool is_ascending = true;
+ bool nulls_first = false;
ndim = ARR_NDIM(array);
dims = ARR_DIMS(array);
@@ -1843,6 +1844,17 @@ array_sort(PG_FUNCTION_ARGS)
if (ndim < 1 || dims[0] < 2)
PG_RETURN_ARRAYTYPE_P(array);
+ if (PG_NARGS() > 1)
+ {
+ is_ascending = PG_GETARG_BOOL(1);
+
+ /*
+ * If nulls_first not provided, it defaults to the opposite of
+ * is_ascending.
+ */
+ nulls_first = PG_NARGS() > 2 ? PG_GETARG_BOOL(2) : !is_ascending;
+ }
+
elmtyp = ARR_ELEMTYPE(array);
cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
if (cache_info == NULL)
@@ -1860,8 +1872,10 @@ array_sort(PG_FUNCTION_ARGS)
typentry = cache_info->typentry;
if (typentry == NULL || typentry->type_id != elmtyp)
{
- typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(elmtyp,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1885,8 +1899,10 @@ array_sort(PG_FUNCTION_ARGS)
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
format_type_be(elmtyp))));
- typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(array_type,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1901,9 +1917,9 @@ array_sort(PG_FUNCTION_ARGS)
}
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
- typentry->lt_opr,
+ is_ascending ? typentry->lt_opr : typentry->gt_opr,
collation,
- false, work_mem, NULL, false);
+ nulls_first, work_mem, NULL, false);
array_iterator = array_create_iterator(array, ndim - 1, &cache_info->array_meta);
while (array_iterate(array_iterator, &value, &isnull))
@@ -1935,3 +1951,15 @@ array_sort(PG_FUNCTION_ARGS)
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
+
+Datum
+array_sort_order_nulls_first(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 54ebbbd135..c6593c63fb 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1747,6 +1747,12 @@
{ oid => '8810', descr => 'sort array',
proname => 'array_sort', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort_order'},
+{ oid => '8812', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool bool', prosrc => 'array_sort_order_nulls_first'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 9e7b444fd8..88c969ccbd 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2771,6 +2771,44 @@ SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo,NULL}
(1 row)
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+ array_sort
+--------------------------------
+ {NULL,1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+ array_sort
+--------------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1,NULL}
+(1 row)
+
-- multidimensional array tests
SELECT array_sort('{{1}}'::int[]);
array_sort
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 2b75a21f06..9cb73f3801 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -843,6 +843,16 @@ SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+
-- multidimensional array tests
SELECT array_sort('{{1}}'::int[]);
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
--
2.39.5
[application/octet-stream] v13-0001-general-purpose-array_sort.patch (15.0K, ../../CAEG8a3KvVjgx5FRWjMq3Jai2LL+xSwDPXsYrMJuL2Jnkr9=zdA@mail.gmail.com/3-v13-0001-general-purpose-array_sort.patch)
download | inline diff:
From 841fcfce37be69758a006d8645ac3ecf48677d18 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v13 1/2] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Author: Junwang Zhao <[email protected]>
Co-authored-by: Jian He <[email protected]>
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 138 ++++++++++++++++++
src/backend/utils/adt/arrayfuncs.c | 3 +-
src/include/catalog/pg_proc.dat | 3 +
src/include/utils/array.h | 1 +
src/test/regress/expected/arrays.out | 86 +++++++++++
.../regress/expected/collate.icu.utf8.out | 13 ++
src/test/regress/sql/arrays.sql | 23 +++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
10 files changed, 289 insertions(+), 1 deletion(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 73979f20ff..748611dee5 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20438,6 +20438,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> )
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
+ The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index d053808f6e..3a50e5ca97 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -42,6 +44,18 @@ typedef struct DeserialIOData
Oid typioparam;
} DeserialIOData;
+/*
+ * ArraySortCachedInfo
+ * Used for caching data in array_sort
+ */
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry; /* type cache entry for element type */
+ TypeCacheEntry *array_typentry; /* type cache entry for array type */
+ ArrayMetaState array_meta; /* array metadata for better
+ * array_create_iterator performance */
+} ArraySortCachedInfo;
+
static Datum array_position_common(FunctionCallInfo fcinfo);
@@ -1797,3 +1811,127 @@ array_reverse(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+ int ndim,
+ *dims,
+ *lbs;
+
+ ndim = ARR_NDIM(array);
+ dims = ARR_DIMS(array);
+ lbs = ARR_LBOUND(array);
+
+ if (ndim < 1 || dims[0] < 2)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ndim == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ cache_info->array_meta.element_type = elmtyp;
+ cache_info->array_meta.typlen = typentry->typlen;
+ cache_info->array_meta.typbyval = typentry->typbyval;
+ cache_info->array_meta.typalign = typentry->typalign;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ cache_info->array_meta.element_type = elmtyp;
+ get_typlenbyvalalign(elmtyp,
+ &cache_info->array_meta.typlen,
+ &cache_info->array_meta.typbyval,
+ &cache_info->array_meta.typalign);
+ }
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ndim - 1, &cache_info->array_meta);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* bounds preservation */
+ if (ndim == 1)
+ astate->scalarstate->lb = lbs[0];
+ else
+ astate->arraystate->lbs[0] = lbs[0];
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/backend/utils/adt/arrayfuncs.c b/src/backend/utils/adt/arrayfuncs.c
index a715e7e0b8..5d5d9b5bbf 100644
--- a/src/backend/utils/adt/arrayfuncs.c
+++ b/src/backend/utils/adt/arrayfuncs.c
@@ -5330,6 +5330,7 @@ initArrayResultWithSize(Oid element_type, MemoryContext rcontext,
MemoryContextAlloc(arr_context, astate->alen * sizeof(bool));
astate->nelems = 0;
astate->element_type = element_type;
+ astate->lb = 1; /* default lower bound */
get_typlenbyvalalign(element_type,
&astate->typlen,
&astate->typbyval,
@@ -5867,7 +5868,7 @@ makeArrayResultAny(ArrayBuildStateAny *astate,
/* If no elements were presented, we want to create an empty array */
ndims = (astate->scalarstate->nelems > 0) ? 1 : 0;
dims[0] = astate->scalarstate->nelems;
- lbs[0] = 1;
+ lbs[0] = astate->scalarstate->lb;
result = makeMdArrayResult(astate->scalarstate, ndims, dims, lbs,
rcontext, release);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index f23321a41f..54ebbbd135 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,9 @@
{ oid => '8686', descr => 'reverse array',
proname => 'array_reverse', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_reverse' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/include/utils/array.h b/src/include/utils/array.h
index 157cc0e4c6..f04cbe4eeb 100644
--- a/src/include/utils/array.h
+++ b/src/include/utils/array.h
@@ -192,6 +192,7 @@ typedef struct ArrayBuildState
int alen; /* allocated length of above arrays */
int nelems; /* number of valid entries in above arrays */
Oid element_type; /* data type of the Datums */
+ int lb; /* lower bound for one dimension array */
int16 typlen; /* needed info about datatype */
bool typbyval;
char typalign;
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 0b61fb5bb7..9e7b444fd8 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2734,3 +2734,89 @@ SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
{{7,8},{5,6},{3,4},{1,2}}
(1 row)
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1}'::int[]);
+ array_sort
+------------
+ {1}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort('{{1}}'::int[]);
+ array_sort
+------------
+ {{1}}
+(1 row)
+
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
+-- no ordering operator tests
+SELECT array_sort('{1,2,3}'::xid[]);
+ERROR: could not identify ordering operator for type xid
+SELECT array_sort('{{1,2,3},{2,3,4}}'::xid[]);
+ERROR: could not identify ordering operator for type xid[]
+-- bounds preservation tests
+SELECT array_sort(a) FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
+ array_sort
+--------------------------------------
+ [10:12][20:21]={{1,2},{3,4},{10,20}}
+(1 row)
+
+SELECT array_sort(a) FROM (VALUES ('[-1:0]={7,1}'::int[])) v(a);
+ array_sort
+--------------
+ [-1:0]={1,7}
+(1 row)
+
+SELECT array_sort(a) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+ array_sort
+--------------------------------------
+ [-2:0][20:21]={{1,-4},{1,2},{10,20}}
+(1 row)
+
+SELECT array_sort(a [-1:0]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+ array_sort
+------------------
+ {{1,-4},{10,20}}
+(1 row)
+
+SELECT array_sort(a [-1:0][20:20]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+ array_sort
+------------
+ {{1},{10}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index faa376e060..aa5fd75e6e 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 03cc8cfcd9..2b75a21f06 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -834,3 +834,26 @@ SELECT array_reverse('{1}'::int[]);
SELECT array_reverse('{1,2}'::int[]);
SELECT array_reverse('{1,2,3,NULL,4,5,6}'::int[]);
SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort('{{1}}'::int[]);
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+
+-- no ordering operator tests
+SELECT array_sort('{1,2,3}'::xid[]);
+SELECT array_sort('{{1,2,3},{2,3,4}}'::xid[]);
+
+-- bounds preservation tests
+SELECT array_sort(a) FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
+SELECT array_sort(a) FROM (VALUES ('[-1:0]={7,1}'::int[])) v(a);
+SELECT array_sort(a) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+SELECT array_sort(a [-1:0]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+SELECT array_sort(a [-1:0][20:20]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 80f28a97d7..3c739d332b 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1847bbfa95..e740e449f5 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-07 14:29 Robert Haas <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 2 replies; 166+ messages in thread
From: Robert Haas @ 2024-11-07 14:29 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: Michael Paquier <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>
On Thu, Nov 7, 2024 at 8:56 AM Junwang Zhao <[email protected]> wrote:
> Yeah, this is reasonable but one case I can't be sure:
>
> SELECT array_sort('{{2,3,4}}'::xid[]);
>
> This will return the array as is, but xid doesn't have a LT_OPR, should
> I error out in this case? like:
>
> could not identify ordering operator for type xid[]
Yes, I think that case needs to error out. It seems best to identify
the ordering operator before you decide whether or not you have >1
element.
--
Robert Haas
EDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-07 14:51 Junwang Zhao <[email protected]>
parent: Robert Haas <[email protected]>
1 sibling, 0 replies; 166+ messages in thread
From: Junwang Zhao @ 2024-11-07 14:51 UTC (permalink / raw)
To: Robert Haas <[email protected]>; +Cc: Michael Paquier <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>
On Thu, Nov 7, 2024 at 10:29 PM Robert Haas <[email protected]> wrote:
>
> On Thu, Nov 7, 2024 at 8:56 AM Junwang Zhao <[email protected]> wrote:
> > Yeah, this is reasonable but one case I can't be sure:
> >
> > SELECT array_sort('{{2,3,4}}'::xid[]);
> >
> > This will return the array as is, but xid doesn't have a LT_OPR, should
> > I error out in this case? like:
> >
> > could not identify ordering operator for type xid[]
>
> Yes, I think that case needs to error out. It seems best to identify
> the ordering operator before you decide whether or not you have >1
> element.
Got it, will do this in the next version.
>
> --
> Robert Haas
> EDB: http://www.enterprisedb.com
--
Regards
Junwang Zhao
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-08 00:01 Michael Paquier <[email protected]>
parent: Robert Haas <[email protected]>
1 sibling, 0 replies; 166+ messages in thread
From: Michael Paquier @ 2024-11-08 00:01 UTC (permalink / raw)
To: Robert Haas <[email protected]>; +Cc: Junwang Zhao <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; jian he <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>
On Thu, Nov 07, 2024 at 09:29:05AM -0500, Robert Haas wrote:
> Yes, I think that case needs to error out. It seems best to identify
> the ordering operator before you decide whether or not you have >1
> element.
+1.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-08 00:51 Michael Paquier <[email protected]>
parent: Junwang Zhao <[email protected]>
0 siblings, 1 reply; 166+ messages in thread
From: Michael Paquier @ 2024-11-08 00:51 UTC (permalink / raw)
To: Junwang Zhao <[email protected]>; +Cc: jian he <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
On Thu, Nov 07, 2024 at 10:06:04PM +0800, Junwang Zhao wrote:
> I'm afraid this can not be achieved in my current implementation, a simple
> case is:
>
> SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[]);
> {Abc,bar,bbc,CCC,foo,NULL}
> SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
> {Abc,CCC,bar,bbc,foo,NULL}
>
> SELECT array_sort(a COLLATE "C") FROM (VALUES
> ('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C"),
> ('{foo,bar,null,CCC,Abc,bbc}'::text[])) v(a);
> {Abc,CCC,bar,bbc,foo,NULL}
> {Abc,CCC,bar,bbc,foo,NULL}
>
> Maybe add some documents to specify this?
So, if I use that:
CREATE COLLATION case_sensitive (provider = icu, locale = '');
=# SELECT array_sort('{Abc,CCC,bar,bbc,foo,NULL}'::text[]
COLLATE "case_sensitive");
array_sort
----------------------------
{Abc,bar,bbc,CCC,foo,NULL}
(1 row)
=# SELECT array_sort('{Abc,CCC,bar,bbc,foo,NULL}'::text[]
COLLATE "C");
array_sort
----------------------------
{Abc,CCC,bar,bbc,foo,NULL}
(1 row)
What takes priority is the collation defined with the array_sort,
which is fine:
=# SELECT array_sort(a collate "case_sensitive") FROM
(VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C" )) v(a);
array_sort
----------------------------
{Abc,bar,bbc,CCC,foo,NULL}
{Abc,bar,bbc,CCC,foo,NULL}
(2 rows)
=# SELECT array_sort(a collate "C") FROM
(VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sensitive" )) v(a);
array_sort
----------------------------
{Abc,CCC,bar,bbc,foo,NULL}
{Abc,CCC,bar,bbc,foo,NULL}
(2 rows)
The case where the collation is defined in the set of values is a bit
more troubling to me, as it depends on what the values want to be
applied, still that's OK because the collation applied is the one
coming from the set of values:
=# SELECT array_sort(a) FROM
(VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sensitive" )) v(a);
array_sort
----------------------------
{Abc,bar,bbc,CCC,foo,NULL}
{Abc,bar,bbc,CCC,foo,NULL}
(2 rows)
=# SELECT array_sort(a) FROM
(VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C" )) v(a);
array_sort
----------------------------
{Abc,CCC,bar,bbc,foo,NULL}
{Abc,CCC,bar,bbc,foo,NULL}
(2 rows)
I am wondering if there are more fancy cases where the saved cache
could force a state that would lead to puzzling results, say with
different collations that should be applied. I'd recommend to
research that more, to reflect that in the docs and to add tests that
show what we should expect in these cases within 0001 because this
new function is mimicking in the context of a function execution
multiple query clauses where restrictions are applied when analyzing
the query, close to the parser.
For example, UNION and UNION ALL require a common collation when
processing a set of expressions related to them, which would be OK.
Perhaps I lack some imagination to be able to break things.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 166+ messages in thread
* Re: general purpose array_sort
@ 2024-11-09 14:45 Junwang Zhao <[email protected]>
parent: Michael Paquier <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Junwang Zhao @ 2024-11-09 14:45 UTC (permalink / raw)
To: Michael Paquier <[email protected]>; +Cc: jian he <[email protected]>; Aleksander Alekseev <[email protected]>; PostgreSQL Hackers <[email protected]>; Tom Lane <[email protected]>; David G. Johnston <[email protected]>; Amit Langote <[email protected]>; [email protected] <[email protected]>; Robert Haas <[email protected]>
Hi Michael,
On Fri, Nov 8, 2024 at 8:52 AM Michael Paquier <[email protected]> wrote:
>
> On Thu, Nov 07, 2024 at 10:06:04PM +0800, Junwang Zhao wrote:
> > I'm afraid this can not be achieved in my current implementation, a simple
> > case is:
> >
> > SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[]);
> > {Abc,bar,bbc,CCC,foo,NULL}
> > SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
> > {Abc,CCC,bar,bbc,foo,NULL}
> >
> > SELECT array_sort(a COLLATE "C") FROM (VALUES
> > ('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C"),
> > ('{foo,bar,null,CCC,Abc,bbc}'::text[])) v(a);
> > {Abc,CCC,bar,bbc,foo,NULL}
> > {Abc,CCC,bar,bbc,foo,NULL}
> >
> > Maybe add some documents to specify this?
>
> So, if I use that:
> CREATE COLLATION case_sensitive (provider = icu, locale = '');
> =# SELECT array_sort('{Abc,CCC,bar,bbc,foo,NULL}'::text[]
> COLLATE "case_sensitive");
> array_sort
> ----------------------------
> {Abc,bar,bbc,CCC,foo,NULL}
> (1 row)
> =# SELECT array_sort('{Abc,CCC,bar,bbc,foo,NULL}'::text[]
> COLLATE "C");
> array_sort
> ----------------------------
> {Abc,CCC,bar,bbc,foo,NULL}
> (1 row)
>
> What takes priority is the collation defined with the array_sort,
> which is fine:
> =# SELECT array_sort(a collate "case_sensitive") FROM
> (VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
> ('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C" )) v(a);
> array_sort
> ----------------------------
> {Abc,bar,bbc,CCC,foo,NULL}
> {Abc,bar,bbc,CCC,foo,NULL}
> (2 rows)
> =# SELECT array_sort(a collate "C") FROM
> (VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
> ('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sensitive" )) v(a);
> array_sort
> ----------------------------
> {Abc,CCC,bar,bbc,foo,NULL}
> {Abc,CCC,bar,bbc,foo,NULL}
> (2 rows)
>
> The case where the collation is defined in the set of values is a bit
> more troubling to me, as it depends on what the values want to be
> applied, still that's OK because the collation applied is the one
> coming from the set of values:
> =# SELECT array_sort(a) FROM
> (VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
> ('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sensitive" )) v(a);
> array_sort
> ----------------------------
> {Abc,bar,bbc,CCC,foo,NULL}
> {Abc,bar,bbc,CCC,foo,NULL}
> (2 rows)
> =# SELECT array_sort(a) FROM
> (VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
> ('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C" )) v(a);
> array_sort
> ----------------------------
> {Abc,CCC,bar,bbc,foo,NULL}
> {Abc,CCC,bar,bbc,foo,NULL}
> (2 rows)
>
> I am wondering if there are more fancy cases where the saved cache
> could force a state that would lead to puzzling results, say with
> different collations that should be applied. I'd recommend to
> research that more, to reflect that in the docs and to add tests that
> show what we should expect in these cases within 0001 because this
> new function is mimicking in the context of a function execution
> multiple query clauses where restrictions are applied when analyzing
> the query, close to the parser.
>
> For example, UNION and UNION ALL require a common collation when
> processing a set of expressions related to them, which would be OK.
> Perhaps I lack some imagination to be able to break things.
> --
> Michael
While trying to come up with more test cases, it comes to me if the
PG_GET_COLLATION() has already done the work to give array_sort
the right collation oid? I did not pass the typentry->typcollation but
PG_GET_COLLATION() to tuplesort_begin_datum.
I tried:
CREATE COLLATION case_sensitive (provider = icu, locale = '');
create table t1(a int, b text[] COLLATE "C");
create table t2(a int, b text[] COLLATE "case_sensitive");
insert into t1 values (1, '{foo,bar,null,CCC,Abc,bbc}'::text[]);
insert into t2 values (2, '{foo,bar,null,CCC,Abc,bbc}'::text[]);
select array_sort(b) from t1;
select array_sort(b) from t2;
Set breakpoint to see the collation oid, they all give the expected result.
For the following cases:
SELECT array_sort(a) FROM
(VALUES ('{foo,bar,null,CCC,Abc,bbc}'::text[]),
('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE
"case_sensitive" )) v(a);
WITH t AS (select '{foo,bar,null,CCC,Abc,bbc}'::text[] a UNION ALL
select '{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sensitive"
a) SELECT array_sort(a) from t;
The collation seems to have been decided in select_common_collation of
the transform phase.
For:
WITH t AS (select '{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C" a
UNION ALL select
'{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sensitive"a)
SELECT array_sort(a) from t;
ERROR: collation mismatch between explicit collations "C" and "case_sensitive"
LINE 2: '{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "case_sens...
merge_collation_state gives out an ERROR since its explicit collation.
But for implicit collation,
select_common_collation sets InvalidOid to context.collation, so the
following works:
select b from t1 UNION ALL select b from t2;
But since the context has the InvalidOid as collation,
PG_GET_COLLATION() in arrary_sort
got InvalidOid, the following errors:
WITH t3 AS (select b from t1 UNION ALL select b from t2) select
array_sort(b) from t3;
ERROR: could not determine which collation to use for string comparison
HINT: Use the COLLATE clause to set the collation explicitly.
The error message comes from tuplesort_begin_datum's call stack, we
can do explicit COLLATE to make it work:
WITH t3 AS (select b from t1 UNION ALL select b from t2) select
array_sort(b collate "C") from t3;
Based on the above analysis, I think it's ok to pass PG_GET_COLLATION()
to tuplesort_begin_datum.
PFA v14 with Robert's comment addressed.
--
Regards
Junwang Zhao
Attachments:
[application/octet-stream] v14-0001-general-purpose-array_sort.patch (15.3K, ../../CAEG8a3KD7ZmQpxNhfPxyc0BjTTTUXiqb56VuMgB7Muu0+yV=qQ@mail.gmail.com/2-v14-0001-general-purpose-array_sort.patch)
download | inline diff:
From 5d810740b0fbfc601a89d1132f064518f0cffc63 Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Fri, 27 Sep 2024 13:05:40 +0000
Subject: [PATCH v14 1/2] general purpose array_sort
Sorts anyarray on its first dimension.
The sort order is determined by the "<" operator.
Author: Junwang Zhao <[email protected]>
Co-authored-by: Jian He <[email protected]>
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 18 +++
src/backend/utils/adt/array_userfuncs.c | 138 ++++++++++++++++++
src/backend/utils/adt/arrayfuncs.c | 3 +-
src/include/catalog/pg_proc.dat | 3 +
src/include/utils/array.h | 1 +
src/test/regress/expected/arrays.out | 90 ++++++++++++
.../regress/expected/collate.icu.utf8.out | 13 ++
src/test/regress/sql/arrays.sql | 25 ++++
src/test/regress/sql/collate.icu.utf8.sql | 4 +
src/tools/pgindent/typedefs.list | 1 +
10 files changed, 295 insertions(+), 1 deletion(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 73979f20ff..748611dee5 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20438,6 +20438,24 @@ SELECT NULLIF(value, '(none)') ...
</para></entry>
</row>
+ <row>
+ <entry role="func_table_entry"><para role="func_signature">
+ <indexterm>
+ <primary>array_sort</primary>
+ </indexterm>
+ <function>array_sort</function> ( <type>anyarray</type> )
+ <returnvalue>anyarray</returnvalue>
+ </para>
+ <para>
+ Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
+ The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ </para>
+ <para>
+ <literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
+ <returnvalue>{{2,1},{2,4},{6,5}}</returnvalue>
+ </para></entry>
+ </row>
+
<row>
<entry role="func_table_entry"><para role="func_signature">
<indexterm id="function-array-to-string">
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index d053808f6e..49b8d5f534 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -16,11 +16,13 @@
#include "common/int.h"
#include "common/pg_prng.h"
#include "libpq/pqformat.h"
+#include "miscadmin.h"
#include "port/pg_bitutils.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
+#include "utils/tuplesort.h"
#include "utils/typcache.h"
/*
@@ -42,6 +44,18 @@ typedef struct DeserialIOData
Oid typioparam;
} DeserialIOData;
+/*
+ * ArraySortCachedInfo
+ * Used for caching data in array_sort
+ */
+typedef struct ArraySortCachedInfo
+{
+ TypeCacheEntry *typentry; /* type cache entry for element type */
+ TypeCacheEntry *array_typentry; /* type cache entry for array type */
+ ArrayMetaState array_meta; /* array metadata for better
+ * array_create_iterator performance */
+} ArraySortCachedInfo;
+
static Datum array_position_common(FunctionCallInfo fcinfo);
@@ -1797,3 +1811,127 @@ array_reverse(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(result);
}
+
+/*
+ * array_sort
+ *
+ * Sorts the first dimension of the array.
+ * The sort order is determined by the "<" operator of the element type.
+ */
+Datum
+array_sort(PG_FUNCTION_ARGS)
+{
+ ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
+ Oid elmtyp;
+ Oid array_type;
+ Oid collation = PG_GET_COLLATION();
+ ArraySortCachedInfo *cache_info;
+ TypeCacheEntry *typentry;
+ Tuplesortstate *tuplesortstate;
+ ArrayIterator array_iterator;
+ Datum value;
+ bool isnull;
+ ArrayBuildStateAny *astate = NULL;
+ int ndim,
+ *dims,
+ *lbs;
+
+ ndim = ARR_NDIM(array);
+ dims = ARR_DIMS(array);
+ lbs = ARR_LBOUND(array);
+
+ elmtyp = ARR_ELEMTYPE(array);
+ cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
+ if (cache_info == NULL)
+ {
+ cache_info = (ArraySortCachedInfo *) MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ sizeof(ArraySortCachedInfo));
+ cache_info->typentry = NULL;
+ cache_info->array_typentry = NULL;
+ fcinfo->flinfo->fn_extra = (void *) cache_info;
+ }
+
+ if (ndim == 1)
+ {
+ /* Finds the ordering operator for the type for 1-D arrays */
+ typentry = cache_info->typentry;
+ if (typentry == NULL || typentry->type_id != elmtyp)
+ {
+ typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(elmtyp))));
+ cache_info->typentry = typentry;
+ cache_info->array_meta.element_type = elmtyp;
+ cache_info->array_meta.typlen = typentry->typlen;
+ cache_info->array_meta.typbyval = typentry->typbyval;
+ cache_info->array_meta.typalign = typentry->typalign;
+ }
+ }
+ else
+ {
+ /* Finds the ordering operator for the array type for multi-D arrays */
+ typentry = cache_info->array_typentry;
+ if (typentry == NULL || typentry->typelem != elmtyp)
+ {
+ array_type = get_array_type(elmtyp);
+ if (!OidIsValid(array_type))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("could not find array type for data type %s",
+ format_type_be(elmtyp))));
+ typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
+ if (!OidIsValid(typentry->lt_opr))
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_FUNCTION),
+ errmsg("could not identify ordering operator for type %s",
+ format_type_be(array_type))));
+ cache_info->array_typentry = typentry;
+ }
+ cache_info->array_meta.element_type = elmtyp;
+ get_typlenbyvalalign(elmtyp,
+ &cache_info->array_meta.typlen,
+ &cache_info->array_meta.typbyval,
+ &cache_info->array_meta.typalign);
+ }
+
+ if (ndim < 1 || dims[0] < 2)
+ PG_RETURN_ARRAYTYPE_P(array);
+
+ tuplesortstate = tuplesort_begin_datum(typentry->type_id,
+ typentry->lt_opr,
+ collation,
+ false, work_mem, NULL, false);
+
+ array_iterator = array_create_iterator(array, ndim - 1, &cache_info->array_meta);
+ while (array_iterate(array_iterator, &value, &isnull))
+ {
+ tuplesort_putdatum(tuplesortstate, value, isnull);
+ }
+ array_free_iterator(array_iterator);
+
+ /*
+ * Do the sort.
+ */
+ tuplesort_performsort(tuplesortstate);
+
+ while (tuplesort_getdatum(tuplesortstate, true, false, &value, &isnull, NULL))
+ {
+ astate = accumArrayResultAny(astate, value, isnull,
+ typentry->type_id, CurrentMemoryContext);
+ }
+
+ tuplesort_end(tuplesortstate);
+
+ /* bounds preservation */
+ if (ndim == 1)
+ astate->scalarstate->lb = lbs[0];
+ else
+ astate->arraystate->lbs[0] = lbs[0];
+
+ /* Avoid leaking memory when handed toasted input */
+ PG_FREE_IF_COPY(array, 0);
+ PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
+}
diff --git a/src/backend/utils/adt/arrayfuncs.c b/src/backend/utils/adt/arrayfuncs.c
index a715e7e0b8..5d5d9b5bbf 100644
--- a/src/backend/utils/adt/arrayfuncs.c
+++ b/src/backend/utils/adt/arrayfuncs.c
@@ -5330,6 +5330,7 @@ initArrayResultWithSize(Oid element_type, MemoryContext rcontext,
MemoryContextAlloc(arr_context, astate->alen * sizeof(bool));
astate->nelems = 0;
astate->element_type = element_type;
+ astate->lb = 1; /* default lower bound */
get_typlenbyvalalign(element_type,
&astate->typlen,
&astate->typbyval,
@@ -5867,7 +5868,7 @@ makeArrayResultAny(ArrayBuildStateAny *astate,
/* If no elements were presented, we want to create an empty array */
ndims = (astate->scalarstate->nelems > 0) ? 1 : 0;
dims[0] = astate->scalarstate->nelems;
- lbs[0] = 1;
+ lbs[0] = astate->scalarstate->lb;
result = makeMdArrayResult(astate->scalarstate, ndims, dims, lbs,
rcontext, release);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index f23321a41f..54ebbbd135 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1744,6 +1744,9 @@
{ oid => '8686', descr => 'reverse array',
proname => 'array_reverse', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_reverse' },
+{ oid => '8810', descr => 'sort array',
+ proname => 'array_sort', prorettype => 'anyarray',
+ proargtypes => 'anyarray', prosrc => 'array_sort'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/include/utils/array.h b/src/include/utils/array.h
index 157cc0e4c6..f04cbe4eeb 100644
--- a/src/include/utils/array.h
+++ b/src/include/utils/array.h
@@ -192,6 +192,7 @@ typedef struct ArrayBuildState
int alen; /* allocated length of above arrays */
int nelems; /* number of valid entries in above arrays */
Oid element_type; /* data type of the Datums */
+ int lb; /* lower bound for one dimension array */
int16 typlen; /* needed info about datatype */
bool typbyval;
char typalign;
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 0b61fb5bb7..9844d5b81f 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2734,3 +2734,93 @@ SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
{{7,8},{5,6},{3,4},{1,2}}
(1 row)
+-- array_sort
+SELECT array_sort('{}'::int[]);
+ array_sort
+------------
+ {}
+(1 row)
+
+SELECT array_sort('{1}'::int[]);
+ array_sort
+------------
+ {1}
+(1 row)
+
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+ array_sort
+---------------
+ {1,2,3,4,5,6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+ array_sort
+---------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+-----------------------
+ {Abc,CCC,bar,bbc,foo}
+(1 row)
+
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+ array_sort
+----------------------------
+ {Abc,CCC,bar,bbc,foo,NULL}
+(1 row)
+
+-- multidimensional array tests
+SELECT array_sort('{{1}}'::int[]);
+ array_sort
+------------
+ {{1}}
+(1 row)
+
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+ array_sort
+---------------------
+ {{2,1},{2,4},{6,5}}
+(1 row)
+
+-- no ordering operator tests
+SELECT array_sort('{1}'::xid[]); -- ndim < 1
+ERROR: could not identify ordering operator for type xid
+SELECT array_sort('{1,2,3}'::xid[]);
+ERROR: could not identify ordering operator for type xid
+SELECT array_sort('{{1,2,3}}'::xid[]); -- dims[0] < 2
+ERROR: could not identify ordering operator for type xid[]
+SELECT array_sort('{{1,2,3},{2,3,4}}'::xid[]);
+ERROR: could not identify ordering operator for type xid[]
+-- bounds preservation tests
+SELECT array_sort(a) FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
+ array_sort
+--------------------------------------
+ [10:12][20:21]={{1,2},{3,4},{10,20}}
+(1 row)
+
+SELECT array_sort(a) FROM (VALUES ('[-1:0]={7,1}'::int[])) v(a);
+ array_sort
+--------------
+ [-1:0]={1,7}
+(1 row)
+
+SELECT array_sort(a) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+ array_sort
+--------------------------------------
+ [-2:0][20:21]={{1,-4},{1,2},{10,20}}
+(1 row)
+
+SELECT array_sort(a [-1:0]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+ array_sort
+------------------
+ {{1,-4},{10,20}}
+(1 row)
+
+SELECT array_sort(a [-1:0][20:20]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+ array_sort
+------------
+ {{1},{10}}
+(1 row)
+
diff --git a/src/test/regress/expected/collate.icu.utf8.out b/src/test/regress/expected/collate.icu.utf8.out
index 6fa32ae364..6d3bbcf1fb 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1338,6 +1338,19 @@ SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_inse
t | t
(1 row)
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+ array_sort
+------------
+ {a,B}
+(1 row)
+
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+ array_sort
+------------
+ {B,a}
+(1 row)
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 03cc8cfcd9..5a5549c8e0 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -834,3 +834,28 @@ SELECT array_reverse('{1}'::int[]);
SELECT array_reverse('{1,2}'::int[]);
SELECT array_reverse('{1,2,3,NULL,4,5,6}'::int[]);
SELECT array_reverse('{{1,2},{3,4},{5,6},{7,8}}'::int[]);
+
+-- array_sort
+SELECT array_sort('{}'::int[]);
+SELECT array_sort('{1}'::int[]);
+SELECT array_sort('{1,3,5,2,4,6}'::int[]);
+SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
+SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
+SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+
+-- multidimensional array tests
+SELECT array_sort('{{1}}'::int[]);
+SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
+
+-- no ordering operator tests
+SELECT array_sort('{1}'::xid[]); -- ndim < 1
+SELECT array_sort('{1,2,3}'::xid[]);
+SELECT array_sort('{{1,2,3}}'::xid[]); -- dims[0] < 2
+SELECT array_sort('{{1,2,3},{2,3,4}}'::xid[]);
+
+-- bounds preservation tests
+SELECT array_sort(a) FROM (VALUES ('[10:12][20:21]={{1,2},{10,20},{3,4}}'::int[])) v(a);
+SELECT array_sort(a) FROM (VALUES ('[-1:0]={7,1}'::int[])) v(a);
+SELECT array_sort(a) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+SELECT array_sort(a [-1:0]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
+SELECT array_sort(a [-1:0][20:20]) FROM (VALUES ('[-2:0][20:21]={{1,2},{10,20},{1,-4}}'::int[])) v(a);
diff --git a/src/test/regress/sql/collate.icu.utf8.sql b/src/test/regress/sql/collate.icu.utf8.sql
index 49fa9758b4..c1e0f6ee1d 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -536,6 +536,10 @@ CREATE COLLATION case_insensitive (provider = icu, locale = '@colStrength=second
SELECT 'abc' <= 'ABC' COLLATE case_sensitive, 'abc' >= 'ABC' COLLATE case_sensitive;
SELECT 'abc' <= 'ABC' COLLATE case_insensitive, 'abc' >= 'ABC' COLLATE case_insensitive;
+-- tests with array_sort
+SELECT array_sort('{a,B}'::text[] COLLATE case_insensitive);
+SELECT array_sort('{a,B}'::text[] COLLATE "C");
+
-- test language tags
CREATE COLLATION lt_insensitive (provider = icu, locale = 'en-u-ks-level1', deterministic = false);
SELECT 'aBcD' COLLATE lt_insensitive = 'AbCd' COLLATE lt_insensitive;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1847bbfa95..e740e449f5 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -150,6 +150,7 @@ ArrayIOData
ArrayIterator
ArrayMapState
ArrayMetaState
+ArraySortCachedInfo
ArraySubWorkspace
ArrayToken
ArrayType
--
2.39.5
[application/octet-stream] v14-0002-support-sort-order-and-nullsfirst-flag.patch (8.7K, ../../CAEG8a3KD7ZmQpxNhfPxyc0BjTTTUXiqb56VuMgB7Muu0+yV=qQ@mail.gmail.com/3-v14-0002-support-sort-order-and-nullsfirst-flag.patch)
download | inline diff:
From 37b20f0a17b8fb54816c319bd474d2906a44b6ba Mon Sep 17 00:00:00 2001
From: Junwang Zhao <[email protected]>
Date: Sat, 26 Oct 2024 03:53:17 +0000
Subject: [PATCH v14 2/2] support sort order and nullsfirst flag
Signed-off-by: Junwang Zhao <[email protected]>
---
doc/src/sgml/func.sgml | 6 +++-
src/backend/utils/adt/array_userfuncs.c | 42 ++++++++++++++++++++-----
src/include/catalog/pg_proc.dat | 6 ++++
src/test/regress/expected/arrays.out | 38 ++++++++++++++++++++++
src/test/regress/sql/arrays.sql | 10 ++++++
5 files changed, 94 insertions(+), 8 deletions(-)
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 748611dee5..afbda78688 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -20443,12 +20443,16 @@ SELECT NULLIF(value, '(none)') ...
<indexterm>
<primary>array_sort</primary>
</indexterm>
- <function>array_sort</function> ( <type>anyarray</type> )
+ <function>array_sort</function> ( <type>anyarray</type> <optional>, <parameter>is_ascending</parameter> <type>boolean</type> <optional>, <parameter>nulls_first</parameter> <type>boolean</type></optional></optional>)
<returnvalue>anyarray</returnvalue>
</para>
<para>
Sorts the first dimension of the array. The sort order is determined by the <literal><</literal> operator of the element type, nulls will appear after non-null values.
The collation to use can be forced by adding a <literal>COLLATE</literal> clause to any of the arguments.
+ If <parameter>is_ascending</parameter> is true then sort by ascending order, otherwise descending order.
+ <parameter>is_ascending</parameter> defaults to true.
+ If <parameter>nulls_first</parameter> is true then nulls appear before non-null values, otherwise nulls appear after non-null values.
+ <parameter>nulls_first</parameter> defaults to the opposite of <parameter>is_ascending</parameter> if not provided.
</para>
<para>
<literal>array_sort(ARRAY[[2,4],[2,1],[6,5]])</literal>
diff --git a/src/backend/utils/adt/array_userfuncs.c b/src/backend/utils/adt/array_userfuncs.c
index 49b8d5f534..33bf6f5243 100644
--- a/src/backend/utils/adt/array_userfuncs.c
+++ b/src/backend/utils/adt/array_userfuncs.c
@@ -1816,7 +1816,6 @@ array_reverse(PG_FUNCTION_ARGS)
* array_sort
*
* Sorts the first dimension of the array.
- * The sort order is determined by the "<" operator of the element type.
*/
Datum
array_sort(PG_FUNCTION_ARGS)
@@ -1835,11 +1834,24 @@ array_sort(PG_FUNCTION_ARGS)
int ndim,
*dims,
*lbs;
+ bool is_ascending = true;
+ bool nulls_first = false;
ndim = ARR_NDIM(array);
dims = ARR_DIMS(array);
lbs = ARR_LBOUND(array);
+ if (PG_NARGS() > 1)
+ {
+ is_ascending = PG_GETARG_BOOL(1);
+
+ /*
+ * If nulls_first not provided, it defaults to the opposite of
+ * is_ascending.
+ */
+ nulls_first = PG_NARGS() > 2 ? PG_GETARG_BOOL(2) : !is_ascending;
+ }
+
elmtyp = ARR_ELEMTYPE(array);
cache_info = (ArraySortCachedInfo *) fcinfo->flinfo->fn_extra;
if (cache_info == NULL)
@@ -1857,8 +1869,10 @@ array_sort(PG_FUNCTION_ARGS)
typentry = cache_info->typentry;
if (typentry == NULL || typentry->type_id != elmtyp)
{
- typentry = lookup_type_cache(elmtyp, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(elmtyp,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1882,8 +1896,10 @@ array_sort(PG_FUNCTION_ARGS)
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("could not find array type for data type %s",
format_type_be(elmtyp))));
- typentry = lookup_type_cache(array_type, TYPECACHE_LT_OPR);
- if (!OidIsValid(typentry->lt_opr))
+ typentry = lookup_type_cache(array_type,
+ is_ascending ? TYPECACHE_LT_OPR : TYPECACHE_GT_OPR);
+ if ((is_ascending && !OidIsValid(typentry->lt_opr)) ||
+ (!is_ascending && !OidIsValid(typentry->gt_opr)))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_FUNCTION),
errmsg("could not identify ordering operator for type %s",
@@ -1901,9 +1917,9 @@ array_sort(PG_FUNCTION_ARGS)
PG_RETURN_ARRAYTYPE_P(array);
tuplesortstate = tuplesort_begin_datum(typentry->type_id,
- typentry->lt_opr,
+ is_ascending ? typentry->lt_opr : typentry->gt_opr,
collation,
- false, work_mem, NULL, false);
+ nulls_first, work_mem, NULL, false);
array_iterator = array_create_iterator(array, ndim - 1, &cache_info->array_meta);
while (array_iterate(array_iterator, &value, &isnull))
@@ -1935,3 +1951,15 @@ array_sort(PG_FUNCTION_ARGS)
PG_FREE_IF_COPY(array, 0);
PG_RETURN_DATUM(makeArrayResultAny(astate, CurrentMemoryContext, true));
}
+
+Datum
+array_sort_order(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
+
+Datum
+array_sort_order_nulls_first(PG_FUNCTION_ARGS)
+{
+ return array_sort(fcinfo);
+}
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 54ebbbd135..c6593c63fb 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -1747,6 +1747,12 @@
{ oid => '8810', descr => 'sort array',
proname => 'array_sort', prorettype => 'anyarray',
proargtypes => 'anyarray', prosrc => 'array_sort'},
+{ oid => '8811', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool', prosrc => 'array_sort_order'},
+{ oid => '8812', descr => 'sort array',
+ proname => 'array_sort', provolatile => 'v', prorettype => 'anyarray',
+ proargtypes => 'anyarray bool bool', prosrc => 'array_sort_order_nulls_first'},
{ oid => '3816', descr => 'array typanalyze',
proname => 'array_typanalyze', provolatile => 's', prorettype => 'bool',
proargtypes => 'internal', prosrc => 'array_typanalyze' },
diff --git a/src/test/regress/expected/arrays.out b/src/test/regress/expected/arrays.out
index 9844d5b81f..3f82972cdf 100644
--- a/src/test/regress/expected/arrays.out
+++ b/src/test/regress/expected/arrays.out
@@ -2771,6 +2771,44 @@ SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
{Abc,CCC,bar,bbc,foo,NULL}
(1 row)
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+ array_sort
+--------------------------------
+ {NULL,1.1,2.2,3.3,4.4,5.5,6.6}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+ array_sort
+--------------------------------
+ {1.1,2.2,3.3,4.4,5.5,6.6,NULL}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+ array_sort
+--------------------------------
+ {NULL,6.6,5.5,4.4,3.3,2.2,1.1}
+(1 row)
+
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+ array_sort
+--------------------------------
+ {6.6,5.5,4.4,3.3,2.2,1.1,NULL}
+(1 row)
+
-- multidimensional array tests
SELECT array_sort('{{1}}'::int[]);
array_sort
diff --git a/src/test/regress/sql/arrays.sql b/src/test/regress/sql/arrays.sql
index 5a5549c8e0..bc19d84e2c 100644
--- a/src/test/regress/sql/arrays.sql
+++ b/src/test/regress/sql/arrays.sql
@@ -843,6 +843,16 @@ SELECT array_sort('{1.1,3.3,5.5,2.2,4.4,6.6}'::numeric[]);
SELECT array_sort('{foo,bar,CCC,Abc,bbc}'::text[] COLLATE "C");
SELECT array_sort('{foo,bar,null,CCC,Abc,bbc}'::text[] COLLATE "C");
+-- array_sort with order specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false);
+
+-- array_sort with order and nullsfirst flag specified
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], true, false);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, true);
+SELECT array_sort('{1.1,3.3,5.5,2.2,null,4.4,6.6}'::float8[], false, false);
+
-- multidimensional array tests
SELECT array_sort('{{1}}'::int[]);
SELECT array_sort(ARRAY[[2,4],[2,1],[6,5]]);
--
2.39.5
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v24 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-12-19 06:06 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-19 06:06 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Dec_19_15_19_50_2024_894)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v24-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v24 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-12-19 06:06 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-19 06:06 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Dec_19_15_19_50_2024_894)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v24-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v24 2/8] Row pattern recognition patch (parse/analysis).
@ 2024-12-19 06:06 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-19 06:06 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Thu_Dec_19_15_19_50_2024_894)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v24-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v25 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-21 06:19 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-21 06:19 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Dec_21_18_20_04_2024_526)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v25-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v25 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-21 06:19 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-21 06:19 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Dec_21_18_20_04_2024_526)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v25-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v25 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-21 06:19 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-21 06:19 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Sat_Dec_21_18_20_04_2024_526)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v25-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v26 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-30 12:44 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-30 12:44 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Dec_30_22_37_18_2024_171)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v26-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v26 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-30 12:44 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-30 12:44 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Dec_30_22_37_18_2024_171)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v26-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v26 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-30 12:44 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-30 12:44 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Mon_Dec_30_22_37_18_2024_171)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v26-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v27 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-30 23:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-30 23:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Dec_31_08_57_07_2024_963)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v27-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v27 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-30 23:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-30 23:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Dec_31_08_57_07_2024_963)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v27-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v27 2/9] Row pattern recognition patch (parse/analysis).
@ 2024-12-30 23:53 Tatsuo Ishii <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Tatsuo Ishii @ 2024-12-30 23:53 UTC (permalink / raw)
---
src/backend/parser/parse_agg.c | 7 +
src/backend/parser/parse_clause.c | 297 +++++++++++++++++++++++++++++-
src/backend/parser/parse_expr.c | 6 +
src/backend/parser/parse_func.c | 3 +
4 files changed, 312 insertions(+), 1 deletion(-)
diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 04b4596a65..6af3bcb375 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -580,6 +580,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
+
/*
* There is intentionally no default: case here, so that the
* compiler will warn if we add a new ParseExprKind without
@@ -970,6 +974,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 979926b605..5a44c68b08 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -96,7 +96,14 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef);
/*
* transformFromClause -
@@ -2954,6 +2961,10 @@ transformWindowDefinitions(ParseState *pstate,
rangeopfamily, rangeopcintype,
&wc->endInRangeFunc,
windef->endOffset);
+
+ /* Process Row Pattern Recognition related clauses */
+ transformRPR(pstate, wc, windef, targetlist);
+
wc->winref = winref;
result = lappend(result, wc);
@@ -3821,3 +3832,287 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
return node;
}
+
+/*
+ * transformRPR
+ * Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /*
+ * Window definition exists?
+ */
+ if (windef == NULL)
+ return;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ /* Check Frame option. Frame must start at current row */
+ if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+ /* Transform AFTER MACH SKIP TO clause */
+ wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+ /* Transform AFTER MACH SKIP TO variable */
+ wc->rpSkipVariable = windef->rpCommonSyntax->rpSkipVariable;
+
+ /* Transform SEEK or INITIAL clause */
+ wc->initial = windef->rpCommonSyntax->initial;
+
+ /* Transform DEFINE clause into list of TargetEntry's */
+ wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+ /* Check PATTERN clause and copy to patternClause */
+ transformPatternClause(pstate, wc, windef);
+
+ /* Transform MEASURE clause */
+ transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause
+ * Process DEFINE clause and transform ResTarget into list of
+ * TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef,
+ List **targetlist)
+{
+ /* DEFINE variable name initials */
+ static char *defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+ ListCell *lc,
+ *l;
+ ResTarget *restarget,
+ *r;
+ List *restargets;
+ List *defineClause;
+ char *name;
+ int initialLen;
+ int i;
+
+ /*
+ * If Row Definition Common Syntax exists, DEFINE clause must exist. (the
+ * raw parser should have already checked it.)
+ */
+ Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+ /*
+ * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+ * per the SQL standard.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ bool found = false;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ foreach(l, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(l);
+
+ if (!strcmp(restarget->name, name))
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ /*
+ * "name" is missing. So create "name AS name IS TRUE" ResTarget
+ * node and add it to the temporary list.
+ */
+ A_Const *n;
+
+ restarget = makeNode(ResTarget);
+ n = makeNode(A_Const);
+ n->val.boolval.type = T_Boolean;
+ n->val.boolval.boolval = true;
+ n->location = -1;
+ restarget->name = pstrdup(name);
+ restarget->indirection = NIL;
+ restarget->val = (Node *) n;
+ restarget->location = -1;
+ restargets = lappend((List *) restargets, restarget);
+ }
+ }
+
+ if (list_length(restargets) >= 1)
+ {
+ /* add missing DEFINEs */
+ windef->rpCommonSyntax->rpDefs =
+ list_concat(windef->rpCommonSyntax->rpDefs, restargets);
+ list_free(restargets);
+ }
+
+ /*
+ * Check for duplicate row pattern definition variables. The standard
+ * requires that no two row pattern definition variable names shall be
+ * equivalent.
+ */
+ restargets = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ /*
+ * Add DEFINE expression (Restarget->val) to the targetlist as a
+ * TargetEntry if it does not exist yet. Planner will add the column
+ * ref var node to the outer plan's target list later on. This makes
+ * DEFINE expression could access the outer tuple while evaluating
+ * PATTERN.
+ *
+ * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+ * not so good, because it's not necessary to evalute the expression
+ * in the target list while running the plan. We should extract the
+ * var nodes only then add them to the plan.targetlist.
+ */
+ findTargetlistEntrySQL99(pstate, (Node *) restarget->val,
+ targetlist, EXPR_KIND_RPR_DEFINE);
+
+ /*
+ * Make sure that the row pattern definition search condition is a
+ * boolean expression.
+ */
+ transformWhereClause(pstate, restarget->val,
+ EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+ foreach(l, restargets)
+ {
+ char *n;
+
+ r = (ResTarget *) lfirst(l);
+ n = r->name;
+
+ if (!strcmp(n, name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+ name),
+ parser_errposition(pstate, exprLocation((Node *) r))));
+ }
+ restargets = lappend(restargets, restarget);
+ }
+ list_free(restargets);
+
+ /*
+ * Create list of row pattern DEFINE variable name's initial. We assign
+ * [a-z] to them (up to 26 variable names are allowed).
+ */
+ restargets = NIL;
+ i = 0;
+ initialLen = strlen(defineVariableInitials);
+
+ foreach(lc, windef->rpCommonSyntax->rpDefs)
+ {
+ char initial[2];
+
+ restarget = (ResTarget *) lfirst(lc);
+ name = restarget->name;
+
+ if (i >= initialLen)
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("number of row pattern definition variable names exceeds %d",
+ initialLen),
+ parser_errposition(pstate,
+ exprLocation((Node *) restarget))));
+ }
+ initial[0] = defineVariableInitials[i++];
+ initial[1] = '\0';
+ wc->defineInitial = lappend(wc->defineInitial,
+ makeString(pstrdup(initial)));
+ }
+
+ defineClause = transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+ EXPR_KIND_RPR_DEFINE);
+
+ /* mark column origins */
+ markTargetListOrigins(pstate, defineClause);
+
+ /* mark all nodes in the DEFINE clause tree with collation information */
+ assign_expr_collations(pstate, (Node *) defineClause);
+
+ return defineClause;
+}
+
+/*
+ * transformPatternClause
+ * Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ ListCell *lc;
+
+ /*
+ * Row Pattern Common Syntax clause exists?
+ */
+ if (windef->rpCommonSyntax == NULL)
+ return;
+
+ wc->patternVariable = NIL;
+ wc->patternRegexp = NIL;
+ foreach(lc, windef->rpCommonSyntax->rpPatterns)
+ {
+ A_Expr *a;
+ char *name;
+ char *regexp;
+
+ if (!IsA(lfirst(lc), A_Expr))
+ ereport(ERROR,
+ errmsg("node type is not A_Expr"));
+
+ a = (A_Expr *) lfirst(lc);
+ name = strVal(a->lexpr);
+
+ wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+ regexp = strVal(lfirst(list_head(a->name)));
+
+ wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+ }
+}
+
+/*
+ * transformMeasureClause
+ * Process MEASURE clause
+ * XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc,
+ WindowDef *windef)
+{
+ if (windef->rowPatternMeasures == NIL)
+ return NIL;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("%s", "MEASURE clause is not supported yet"),
+ parser_errposition(pstate, exprLocation((Node *) windef->rowPatternMeasures))));
+ return NIL;
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index c2806297aa..e827c59fd4 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -575,6 +575,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
case EXPR_KIND_COPY_WHERE:
case EXPR_KIND_GENERATED_COLUMN:
case EXPR_KIND_CYCLE_MARK:
+ case EXPR_KIND_RPR_DEFINE:
/* okay */
break;
@@ -1858,6 +1859,9 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
case EXPR_KIND_GENERATED_COLUMN:
err = _("cannot use subquery in column generation expression");
break;
+ case EXPR_KIND_RPR_DEFINE:
+ err = _("cannot use subquery in DEFINE expression");
+ break;
/*
* There is intentionally no default: case here, so that the
@@ -3197,6 +3201,8 @@ ParseExprKindName(ParseExprKind exprKind)
return "GENERATED AS";
case EXPR_KIND_CYCLE_MARK:
return "CYCLE";
+ case EXPR_KIND_RPR_DEFINE:
+ return "DEFINE";
/*
* There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index 9b23344a3b..4c482abb30 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2658,6 +2658,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
case EXPR_KIND_CYCLE_MARK:
errkind = true;
break;
+ case EXPR_KIND_RPR_DEFINE:
+ errkind = true;
+ break;
/*
* There is intentionally no default: case here, so that the
--
2.25.1
----Next_Part(Tue_Dec_31_08_57_07_2024_963)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="v27-0003-Row-pattern-recognition-patch-rewriter.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushBehavior enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- Register ANYTIME_STATS_UPDATE_TIMEOUT that fires every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/tcop/postgres.c | 18 +++++
src/backend/utils/activity/pgstat.c | 119 ++++++++++++++++++++++++----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 11 +++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
9 files changed, 154 insertions(+), 17 deletions(-)
9.7% src/backend/tcop/
70.2% src/backend/utils/activity/
9.3% src/backend/utils/init/
6.0% src/include/utils/
4.3% src/include/
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..6a91543f80a 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,24 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ /* Skip if completely idle */
+ if (!DoingCommandRead || IsTransactionOrTransactionBlock())
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_ANYTIME_FLUSH_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..f7942e47475 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +289,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +307,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +324,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +340,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +358,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +376,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_behavior = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +395,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +412,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +429,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +446,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +464,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +482,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +788,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1346,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1383,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1397,11 +1417,42 @@ pgstat_flush_pending_entries(bool nowait)
cur = next;
}
- Assert(dlist_is_empty(&pgStatPending) == !have_pending);
+ /*
+ * When in anytime_only mode, the list may not be empty because
+ * FLUSH_AT_TXN_BOUNDARY entries were skipped.
+ */
+ Assert(!anytime_only || dlist_is_empty(&pgStatPending) == !have_pending);
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,37 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flush non-transactional stats
+ *
+ * This is safe to call even inside a transaction. It only flushes stats
+ * kinds marked as FLUSH_ANYTIME.
+ *
+ * This allows long running transactions to report activity without waiting
+ * for transaction to finish.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..cb0f6aecad1 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_ANYTIME_FLUSH_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..86e65397614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* When to call pgstat_report_anytime_stat() again */
+#define PGSTAT_ANYTIME_FLUSH_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..02f4f13fc0f 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,14 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush behavior for statistics kinds.
+ */
+typedef enum PgStat_FlushBehavior
+{
+ FLUSH_ANYTIME, /* All fields can flush anytime */
+ FLUSH_AT_TXN_BOUNDARY, /* All fields need transaction boundary */
+} PgStat_FlushBehavior;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +259,9 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /* Flush behavior */
+ PgStat_FlushBehavior flush_behavior;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 09e7f1d420e..9aabb325f16 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2261,6 +2261,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushBehavior
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--tjEWjIIwfNIHQLgt
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v1-0002-Remove-useless-calls-to-flush-some-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v11 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 121 +++++++++++++++---
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 22 ++++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 279 insertions(+), 66 deletions(-)
10.5% src/backend/postmaster/
5.8% src/backend/replication/
51.0% src/backend/utils/activity/
5.8% src/backend/
18.7% src/include/utils/
6.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13cce9b49f1..cf29fc91f70 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 742137edad6..f1bae9d23d6 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -246,7 +248,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -268,6 +270,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 7c0e2809c17..bcf59227a00 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -103,7 +105,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -113,6 +115,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 7c1b8757d7d..aecc7a127e6 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index d01a09dd0c4..8c30efa2443 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ddd331e2c81 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -108,10 +108,12 @@
#include "pgstat.h"
#include "storage/fd.h"
#include "storage/ipc.h"
+#include "storage/latch.h"
#include "storage/lwlock.h"
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +124,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -218,6 +219,12 @@ PgStat_LocalState pgStatLocal;
*/
bool pgstat_report_fixed = false;
+/*
+ * Track when there is pending anytime flush to avoid relying on
+ * get_timeout_active() in hot pathes.
+ */
+bool pgstat_pending_anytime = false;
+
/* ----------
* Local data
*
@@ -288,6 +295,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +313,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +330,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +346,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +364,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +382,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +449,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +467,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +485,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +791,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1297,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1350,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1387,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1426,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+
+ pgstat_pending_anytime = false;
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 3277cf88a4e..6b6eec7578d 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -117,11 +117,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 9bb777c3d5a..b011a315679 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -34,6 +34,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -532,8 +535,24 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !pgstat_pending_anytime) \
+ { \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ pgstat_pending_anytime = true; \
+ } \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
@@ -806,6 +825,8 @@ extern PgStat_WalStats *pgstat_fetch_stat_wal(void);
* Variables in pgstat.c
*/
+extern PGDLLIMPORT bool pgstat_pending_anytime;
+
/* GUC parameters */
extern PGDLLIMPORT bool pgstat_track_counts;
extern PGDLLIMPORT int pgstat_track_functions;
@@ -849,4 +870,5 @@ extern PGDLLIMPORT PgStat_Counter pgStatTransactionIdleTime;
/* updated by the traffic cop and in errfinish() */
extern PGDLLIMPORT SessionEndType pgStatSessionEndCause;
+
#endif /* PGSTAT_H */
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index da28afbd929..4c207611236 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -84,7 +84,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -151,7 +151,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--2MEBAGW8+kohXisi
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v11-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v10 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 121 +++++++++++++++---
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 22 ++++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 279 insertions(+), 66 deletions(-)
10.5% src/backend/postmaster/
5.8% src/backend/replication/
51.0% src/backend/utils/activity/
5.8% src/backend/
18.7% src/include/utils/
6.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13cce9b49f1..cf29fc91f70 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 742137edad6..f1bae9d23d6 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -246,7 +248,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -268,6 +270,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 7c0e2809c17..bcf59227a00 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -103,7 +105,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -113,6 +115,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 7c1b8757d7d..aecc7a127e6 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index d01a09dd0c4..8c30efa2443 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ddd331e2c81 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -108,10 +108,12 @@
#include "pgstat.h"
#include "storage/fd.h"
#include "storage/ipc.h"
+#include "storage/latch.h"
#include "storage/lwlock.h"
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +124,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -218,6 +219,12 @@ PgStat_LocalState pgStatLocal;
*/
bool pgstat_report_fixed = false;
+/*
+ * Track when there is pending anytime flush to avoid relying on
+ * get_timeout_active() in hot pathes.
+ */
+bool pgstat_pending_anytime = false;
+
/* ----------
* Local data
*
@@ -288,6 +295,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +313,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +330,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +346,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +364,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +382,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +449,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +467,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +485,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +791,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1297,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1350,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1387,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1426,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+
+ pgstat_pending_anytime = false;
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 3277cf88a4e..6b6eec7578d 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -117,11 +117,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 9bb777c3d5a..b011a315679 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -34,6 +34,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -532,8 +535,24 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !pgstat_pending_anytime) \
+ { \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ pgstat_pending_anytime = true; \
+ } \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
@@ -806,6 +825,8 @@ extern PgStat_WalStats *pgstat_fetch_stat_wal(void);
* Variables in pgstat.c
*/
+extern PGDLLIMPORT bool pgstat_pending_anytime;
+
/* GUC parameters */
extern PGDLLIMPORT bool pgstat_track_counts;
extern PGDLLIMPORT int pgstat_track_functions;
@@ -849,4 +870,5 @@ extern PGDLLIMPORT PgStat_Counter pgStatTransactionIdleTime;
/* updated by the traffic cop and in errfinish() */
extern PGDLLIMPORT SessionEndType pgStatSessionEndCause;
+
#endif /* PGSTAT_H */
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index da28afbd929..4c207611236 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -84,7 +84,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -151,7 +151,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--NVvBxFuyV/R+1/8/
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v10-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v9 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 120 +++++++++++++++---
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 22 ++++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 278 insertions(+), 66 deletions(-)
10.5% src/backend/postmaster/
5.8% src/backend/replication/
50.9% src/backend/utils/activity/
5.9% src/backend/
18.8% src/include/utils/
6.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..d01b11c7470 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 742137edad6..f1bae9d23d6 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -246,7 +248,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -268,6 +270,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 7c0e2809c17..bcf59227a00 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -103,7 +105,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -113,6 +115,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 10e64a7d1f4..11b7c114d3b 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 21de158adbb..2089de782d5 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..419dc512d9b 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -112,6 +112,7 @@
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +123,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +186,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -218,6 +218,12 @@ PgStat_LocalState pgStatLocal;
*/
bool pgstat_report_fixed = false;
+/*
+ * Track when there is pending anytime flush to avoid relying on
+ * get_timeout_active() in hot pathes.
+ */
+bool pgstat_pending_anytime = false;
+
/* ----------
* Local data
*
@@ -288,6 +294,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +312,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +329,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +345,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +363,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +381,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +448,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +466,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +484,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +790,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1296,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1349,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1386,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1425,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2169,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+
+ pgstat_pending_anytime = false;
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 500b1899188..c4614817966 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -116,11 +116,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..f0f546d419a 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,8 +536,24 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !pgstat_pending_anytime) \
+ { \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ pgstat_pending_anytime = true; \
+ } \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
@@ -808,6 +827,8 @@ extern PgStat_WalStats *pgstat_fetch_stat_wal(void);
* Variables in pgstat.c
*/
+extern PGDLLIMPORT bool pgstat_pending_anytime;
+
/* GUC parameters */
extern PGDLLIMPORT bool pgstat_track_counts;
extern PGDLLIMPORT int pgstat_track_functions;
@@ -851,4 +872,5 @@ extern PGDLLIMPORT PgStat_Counter pgStatTransactionIdleTime;
/* updated by the traffic cop and in errfinish() */
extern PGDLLIMPORT SessionEndType pgStatSessionEndCause;
+
#endif /* PGSTAT_H */
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index 64a8fe63cce..bc0b5d6e0eb 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -83,7 +83,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -150,7 +150,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--aq/6bi8L6WORnI+9
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v9-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v8 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 112 ++++++++++++++----
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 16 +++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 264 insertions(+), 66 deletions(-)
10.8% src/backend/postmaster/
6.0% src/backend/replication/
50.7% src/backend/utils/activity/
6.0% src/backend/
19.3% src/include/utils/
5.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..d01b11c7470 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 742137edad6..f1bae9d23d6 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -246,7 +248,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -268,6 +270,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 7c0e2809c17..bcf59227a00 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -103,7 +105,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -113,6 +115,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 10e64a7d1f4..11b7c114d3b 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 21de158adbb..2089de782d5 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..a4ff64dc5ce 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -112,6 +112,7 @@
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +123,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +186,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +288,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +306,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +323,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +339,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +357,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +375,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +442,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +460,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +478,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +784,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1290,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1343,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1380,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1419,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2163,31 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 500b1899188..c4614817966 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -116,11 +116,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..b340a680614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,8 +536,21 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT)) \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index 64a8fe63cce..bc0b5d6e0eb 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -83,7 +83,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -150,7 +150,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--OI5irqZsxWxBW9nT
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v8-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 112 ++++++++++++++----
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 16 +++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 264 insertions(+), 66 deletions(-)
10.8% src/backend/postmaster/
6.0% src/backend/replication/
50.7% src/backend/utils/activity/
6.0% src/backend/
19.3% src/include/utils/
5.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..d01b11c7470 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 2d8f57099fd..9f8ef8159d1 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -249,7 +251,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -271,6 +273,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 23e79a32345..ded0f250288 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -106,7 +108,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -116,6 +118,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 10e64a7d1f4..11b7c114d3b 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 21de158adbb..2089de782d5 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..a4ff64dc5ce 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -112,6 +112,7 @@
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +123,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +186,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +288,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +306,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +323,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +339,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +357,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +375,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +442,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +460,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +478,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +784,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1290,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1343,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1380,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1419,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2163,31 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 500b1899188..c4614817966 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -116,11 +116,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..b340a680614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,8 +536,21 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT)) \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index 64a8fe63cce..bc0b5d6e0eb 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -83,7 +83,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -150,7 +150,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--C2xzVbxFmVrP7k6O
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v7-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v6 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 112 ++++++++++++++----
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 16 +++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 264 insertions(+), 66 deletions(-)
10.8% src/backend/postmaster/
6.0% src/backend/replication/
50.7% src/backend/utils/activity/
6.0% src/backend/
19.3% src/include/utils/
5.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..d01b11c7470 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 2d8f57099fd..9f8ef8159d1 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -249,7 +251,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -271,6 +273,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 23e79a32345..ded0f250288 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -106,7 +108,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -116,6 +118,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 10e64a7d1f4..11b7c114d3b 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 02e9aaa6bca..c7bc409b06f 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3509,6 +3509,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..411b65aae3e 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -112,6 +112,7 @@
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +123,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +186,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +288,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +306,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +323,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +339,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +357,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +375,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +442,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +460,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +478,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +784,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1290,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1343,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1380,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1419,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2163,31 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index 1350f5f62f1..b0bd220882d 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index d7f6d4c5ee6..61094f96a6c 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -425,7 +425,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 500b1899188..c4614817966 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -116,11 +116,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..eb5f8f46925 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..f45365f47f7 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -765,6 +766,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..b340a680614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,8 +536,21 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT)) \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index 64a8fe63cce..bc0b5d6e0eb 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -83,7 +83,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -150,7 +150,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 9f5ee8fd482..860f835c088 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--DxqFthphbM+ha9Dy
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v6-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v5 1/4] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- This relies on the existing PGSTAT_MIN_INTERVAL to fire every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 9 ++
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 118 ++++++++++++++++----
src/backend/utils/activity/pgstat_backend.c | 9 ++
src/backend/utils/activity/pgstat_io.c | 5 +
src/backend/utils/activity/pgstat_slru.c | 5 +
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 21 ++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
19 files changed, 213 insertions(+), 25 deletions(-)
5.5% src/backend/access/transam/
15.3% src/backend/postmaster/
3.4% src/backend/replication/
3.8% src/backend/tcop/
57.0% src/backend/utils/activity/
9.2% src/include/utils/
5.3% src/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..9503aea5b4d 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,10 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2070,11 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_MIN_INTERVAL);
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 80e3088fc7e..ab5d0645026 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -104,7 +106,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -114,6 +116,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index 6482c21b8f9..6e187315613 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -216,7 +217,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -226,6 +227,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index a1a4f65f9a9..498d147f0da 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -246,6 +247,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index c3d56c866d3..cec5dfdb430 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -250,7 +252,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -272,6 +274,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 38ec8a4c8c7..7416ca703c9 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -107,7 +109,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -117,6 +119,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 8b99160ed0e..24d7ef795cb 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -253,7 +255,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -261,6 +263,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index b4a8d2f3a1c..d19aa45400d 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..2c9454677e9 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -112,6 +112,7 @@
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +123,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +186,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +288,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +306,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +323,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +339,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +357,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +375,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +394,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +411,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +428,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +445,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +463,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +481,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +787,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,12 +1293,18 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
entry_ref->pending = MemoryContextAllocZero(pgStatPendingContext, entrysize);
dlist_push_tail(&pgStatPending, &entry_ref->pending_node);
+
+ /* Schedule next anytime stats update timeout */
+ if (kind_info->flush_mode == FLUSH_ANYTIME && IsUnderPostmaster &&
+ !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
}
return entry_ref;
@@ -1345,9 +1351,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1388,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1402,6 +1427,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2171,31 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index 1350f5f62f1..9dcb24db975 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,10 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +87,10 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..53dbf2a514b 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,10 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
have_iostats = true;
pgstat_report_fixed = true;
}
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..1d16cde1889 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -223,6 +224,10 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT))
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..f45365f47f7 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -765,6 +766,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..1651f16f966 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..a9190078d0e 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,13 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * Some stats have to be updated only at transaction boundaries (such as
+ * tuples_inserted updated, deleted), so it's very important to set the
+ * right flush mode (FLUSH_AT_TXN_BOUNDARY being the default).
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -677,6 +697,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 9f5ee8fd482..860f835c088 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--WLtTrIUEfQDmERxy
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v5-0002-Add-GUC-to-specify-non-transactional-statistics-f.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 1/4] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- This relies on the existing PGSTAT_MIN_INTERVAL to fire every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/storage/lmgr/proc.c | 10 +++
src/backend/tcop/postgres.c | 16 ++++
src/backend/utils/activity/pgstat.c | 111 +++++++++++++++++++++++-----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 20 +++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
10 files changed, 162 insertions(+), 18 deletions(-)
7.0% src/backend/storage/lmgr/
7.3% src/backend/tcop/
61.1% src/backend/utils/activity/
8.6% src/backend/utils/init/
11.8% src/include/utils/
3.6% src/include/
diff --git a/src/backend/storage/lmgr/proc.c b/src/backend/storage/lmgr/proc.c
index 063826ae576..012705a2ee6 100644
--- a/src/backend/storage/lmgr/proc.c
+++ b/src/backend/storage/lmgr/proc.c
@@ -1322,6 +1322,7 @@ ProcSleep(LOCALLOCK *locallock)
bool allow_autovacuum_cancel = true;
bool logged_recovery_conflict = false;
ProcWaitStatus myWaitStatus;
+ bool anytime_timeout_was_active = false;
/* The caller must've armed the on-error cleanup mechanism */
Assert(GetAwaitedLock() == locallock);
@@ -1398,6 +1399,12 @@ ProcSleep(LOCALLOCK *locallock)
standbyWaitStart = GetCurrentTimestamp();
}
+ anytime_timeout_was_active = get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT);
+
+ /* No need to try to flush the statistics while the process is sleeping */
+ if (anytime_timeout_was_active)
+ disable_timeout(ANYTIME_STATS_UPDATE_TIMEOUT, false);
+
/*
* If somebody wakes us between LWLockRelease and WaitLatch, the latch
* will not wait. But a set latch does not necessarily mean that the lock
@@ -1661,6 +1668,9 @@ ProcSleep(LOCALLOCK *locallock)
}
} while (myWaitStatus == PROC_WAIT_STATUS_WAITING);
+ if (anytime_timeout_was_active)
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
/*
* Disable the timers, if they are still running. As in LockErrorCleanup,
* we must preserve the LOCK_TIMEOUT indicator flag: if a lock timeout has
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..132fae61423 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,22 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_MIN_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ab4d9088a9a 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -122,8 +122,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +185,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +287,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +305,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +322,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +338,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +356,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +374,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +393,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +410,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +427,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +444,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +462,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +480,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +786,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1344,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1381,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1402,6 +1420,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2164,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..6076f531c4a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..1651f16f966 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..9ca39ea9a9a 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,13 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * Some stats have to be updated only at transaction boundaries (such as
+ * tuples_inserted updated, deleted), so it's very important to set the
+ * right flush mode (FLUSH_AT_TXN_BOUNDARY being the default).
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ddbe4c64971..af21c87234a 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--wR/mWGXukst4NraF
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v4-0002-Add-GUC-to-specify-non-transactional-statistics-f.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v3 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- This relies on the existing PGSTAT_MIN_INTERVAL to fire every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/tcop/postgres.c | 16 ++++
src/backend/utils/activity/pgstat.c | 111 +++++++++++++++++++++++-----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 16 ++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
9 files changed, 148 insertions(+), 18 deletions(-)
8.1% src/backend/tcop/
68.4% src/backend/utils/activity/
9.6% src/backend/utils/init/
9.2% src/include/utils/
4.1% src/include/
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..132fae61423 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,22 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_MIN_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ab4d9088a9a 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -122,8 +122,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +185,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +287,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +305,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +322,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +338,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +356,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +374,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +393,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +410,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +427,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +444,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +462,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +480,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +786,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1344,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1381,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1402,6 +1420,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2164,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..6076f531c4a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..1651f16f966 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..46ce90c9624 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,9 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /* Flush mode */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 3f3a888fd0e..d3912b43fdc 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--lCc1MWfC7i2uqs8g
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v3-0002-Remove-useless-calls-to-flush-some-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushBehavior enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- Register ANYTIME_STATS_UPDATE_TIMEOUT that fires every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/tcop/postgres.c | 16 ++++
src/backend/utils/activity/pgstat.c | 113 ++++++++++++++++++++++++----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 13 ++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
9 files changed, 149 insertions(+), 16 deletions(-)
8.2% src/backend/tcop/
69.3% src/backend/utils/activity/
9.7% src/backend/utils/init/
7.7% src/include/utils/
4.5% src/include/
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..9c4a9078ee0 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,22 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_ANYTIME_FLUSH_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..0f45a7d165e 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +289,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +307,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +324,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +340,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +358,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +376,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_behavior = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +395,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +412,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +429,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +446,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +464,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +482,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +788,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1346,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1383,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1397,11 +1417,42 @@ pgstat_flush_pending_entries(bool nowait)
cur = next;
}
+ /*
+ * When in anytime_only mode, the list may not be empty because
+ * FLUSH_AT_TXN_BOUNDARY entries were skipped.
+ */
Assert(dlist_is_empty(&pgStatPending) == !have_pending);
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..cb0f6aecad1 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_ANYTIME_FLUSH_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..86e65397614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* When to call pgstat_report_anytime_stat() again */
+#define PGSTAT_ANYTIME_FLUSH_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..63feae640d1 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,16 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush behavior for statistics kinds.
+ */
+typedef enum PgStat_FlushBehavior
+{
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+} PgStat_FlushBehavior;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +261,9 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /* Flush behavior */
+ PgStat_FlushBehavior flush_behavior;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 3f3a888fd0e..610b35a9b31 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushBehavior
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--OFsrIl+bjhifp5hk
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v2-0002-Remove-useless-calls-to-flush-some-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v1 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushBehavior enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- Register ANYTIME_STATS_UPDATE_TIMEOUT that fires every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/tcop/postgres.c | 18 +++++
src/backend/utils/activity/pgstat.c | 119 ++++++++++++++++++++++++----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 11 +++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
9 files changed, 154 insertions(+), 17 deletions(-)
9.7% src/backend/tcop/
70.2% src/backend/utils/activity/
9.3% src/backend/utils/init/
6.0% src/include/utils/
4.3% src/include/
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..6a91543f80a 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,24 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ /* Skip if completely idle */
+ if (!DoingCommandRead || IsTransactionOrTransactionBlock())
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_ANYTIME_FLUSH_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..f7942e47475 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +289,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +307,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +324,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +340,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +358,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +376,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_behavior = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +395,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +412,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +429,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +446,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +464,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +482,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +788,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1346,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1383,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1397,11 +1417,42 @@ pgstat_flush_pending_entries(bool nowait)
cur = next;
}
- Assert(dlist_is_empty(&pgStatPending) == !have_pending);
+ /*
+ * When in anytime_only mode, the list may not be empty because
+ * FLUSH_AT_TXN_BOUNDARY entries were skipped.
+ */
+ Assert(!anytime_only || dlist_is_empty(&pgStatPending) == !have_pending);
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,37 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flush non-transactional stats
+ *
+ * This is safe to call even inside a transaction. It only flushes stats
+ * kinds marked as FLUSH_ANYTIME.
+ *
+ * This allows long running transactions to report activity without waiting
+ * for transaction to finish.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..cb0f6aecad1 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_ANYTIME_FLUSH_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..86e65397614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* When to call pgstat_report_anytime_stat() again */
+#define PGSTAT_ANYTIME_FLUSH_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..02f4f13fc0f 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,14 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush behavior for statistics kinds.
+ */
+typedef enum PgStat_FlushBehavior
+{
+ FLUSH_ANYTIME, /* All fields can flush anytime */
+ FLUSH_AT_TXN_BOUNDARY, /* All fields need transaction boundary */
+} PgStat_FlushBehavior;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +259,9 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /* Flush behavior */
+ PgStat_FlushBehavior flush_behavior;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 09e7f1d420e..9aabb325f16 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2261,6 +2261,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushBehavior
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--tjEWjIIwfNIHQLgt
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v1-0002-Remove-useless-calls-to-flush-some-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v11 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 121 +++++++++++++++---
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 22 ++++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 279 insertions(+), 66 deletions(-)
10.5% src/backend/postmaster/
5.8% src/backend/replication/
51.0% src/backend/utils/activity/
5.8% src/backend/
18.7% src/include/utils/
6.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13cce9b49f1..cf29fc91f70 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 742137edad6..f1bae9d23d6 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -246,7 +248,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -268,6 +270,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 7c0e2809c17..bcf59227a00 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -103,7 +105,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN); /* no query to cancel */
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -113,6 +115,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 7c1b8757d7d..aecc7a127e6 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index d01a09dd0c4..8c30efa2443 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ddd331e2c81 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -108,10 +108,12 @@
#include "pgstat.h"
#include "storage/fd.h"
#include "storage/ipc.h"
+#include "storage/latch.h"
#include "storage/lwlock.h"
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +124,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -218,6 +219,12 @@ PgStat_LocalState pgStatLocal;
*/
bool pgstat_report_fixed = false;
+/*
+ * Track when there is pending anytime flush to avoid relying on
+ * get_timeout_active() in hot pathes.
+ */
+bool pgstat_pending_anytime = false;
+
/* ----------
* Local data
*
@@ -288,6 +295,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +313,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +330,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +346,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +364,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +382,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +449,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +467,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +485,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +791,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1297,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1350,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1387,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1426,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+
+ pgstat_pending_anytime = false;
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 3277cf88a4e..6b6eec7578d 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -117,11 +117,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 9bb777c3d5a..b011a315679 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -34,6 +34,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -532,8 +535,24 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !pgstat_pending_anytime) \
+ { \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ pgstat_pending_anytime = true; \
+ } \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
@@ -806,6 +825,8 @@ extern PgStat_WalStats *pgstat_fetch_stat_wal(void);
* Variables in pgstat.c
*/
+extern PGDLLIMPORT bool pgstat_pending_anytime;
+
/* GUC parameters */
extern PGDLLIMPORT bool pgstat_track_counts;
extern PGDLLIMPORT int pgstat_track_functions;
@@ -849,4 +870,5 @@ extern PGDLLIMPORT PgStat_Counter pgStatTransactionIdleTime;
/* updated by the traffic cop and in errfinish() */
extern PGDLLIMPORT SessionEndType pgStatSessionEndCause;
+
#endif /* PGSTAT_H */
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index da28afbd929..4c207611236 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -84,7 +84,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -151,7 +151,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--2MEBAGW8+kohXisi
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v11-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v7 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal and produces
spikes when flushed.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second (if enabled while adding pending stats) to call this
function, ensuring timely stats visibility without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- The flush_pending_cb and flush_static_cb callbacks now receive an anytime_only
boolean parameter. Most of the time it's not used (except for assertions), but it's
preparatory work for moving the relations stats to anytime (without introducin
a new callback).
- Add pgstat_schedule_anytime_update() macro to schedule the next anytime flush,
relying on PGSTAT_MIN_INTERVAL
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/access/transam/xlog.c | 6 +
src/backend/postmaster/bgwriter.c | 9 +-
src/backend/postmaster/checkpointer.c | 10 +-
src/backend/postmaster/startup.c | 2 +
src/backend/postmaster/walsummarizer.c | 9 +-
src/backend/postmaster/walwriter.c | 9 +-
src/backend/replication/walreceiver.c | 9 +-
src/backend/replication/walsender.c | 8 +-
src/backend/tcop/postgres.c | 12 ++
src/backend/utils/activity/pgstat.c | 112 ++++++++++++++----
src/backend/utils/activity/pgstat_backend.c | 13 +-
src/backend/utils/activity/pgstat_bgwriter.c | 2 +-
.../utils/activity/pgstat_checkpointer.c | 2 +-
src/backend/utils/activity/pgstat_database.c | 2 +-
src/backend/utils/activity/pgstat_function.c | 4 +-
src/backend/utils/activity/pgstat_io.c | 10 +-
src/backend/utils/activity/pgstat_relation.c | 12 +-
src/backend/utils/activity/pgstat_slru.c | 6 +-
.../utils/activity/pgstat_subscription.c | 4 +-
src/backend/utils/activity/pgstat_wal.c | 10 +-
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 3 +
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 16 +++
src/include/utils/pgstat_internal.h | 52 ++++++--
src/include/utils/timeout.h | 1 +
.../test_custom_stats/test_custom_var_stats.c | 4 +-
src/tools/pgindent/typedefs.list | 1 +
28 files changed, 264 insertions(+), 66 deletions(-)
10.8% src/backend/postmaster/
6.0% src/backend/replication/
50.7% src/backend/utils/activity/
6.0% src/backend/
19.3% src/include/utils/
5.6% src/include/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..d01b11c7470 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -1085,6 +1085,9 @@ XLogInsertRecord(XLogRecData *rdata,
pgWalUsage.wal_fpi += num_fpi;
pgWalUsage.wal_fpi_bytes += fpi_bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/* Required for the flush of pending stats WAL data */
pgstat_report_fixed = true;
}
@@ -2066,6 +2069,9 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic)
pgWalUsage.wal_buffers_full++;
TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE();
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
/*
* Required for the flush of pending stats WAL data, per
* update of pgWalUsage.
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 0956bd39a85..059c601c3b8 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -49,7 +49,9 @@
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
/*
@@ -103,7 +105,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -113,6 +115,11 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* We just started, assume there has been either a shutdown or
* end-of-recovery snapshot.
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index e03c19123bc..e11c4b099c8 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -66,8 +66,9 @@
#include "utils/acl.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
-
+#include "utils/timeout.h"
/*----------
* Shared memory area for communication between checkpointer and backends
@@ -215,7 +216,7 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, ReqShutdownXLOG);
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SignalHandlerForShutdownRequest);
@@ -225,6 +226,11 @@ CheckpointerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Initialize so that first time-driven event happens at the correct time.
*/
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index cdbe53dd262..4954fe425b7 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -32,6 +32,7 @@
#include "storage/standby.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/timeout.h"
@@ -245,6 +246,7 @@ StartupProcessMain(const void *startup_data, size_t startup_data_len)
RegisterTimeout(STANDBY_DEADLOCK_TIMEOUT, StandbyDeadLockHandler);
RegisterTimeout(STANDBY_TIMEOUT, StandbyTimeoutHandler);
RegisterTimeout(STANDBY_LOCK_TIMEOUT, StandbyLockTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
/*
* Unblock signals (they were blocked when the postmaster forked us)
diff --git a/src/backend/postmaster/walsummarizer.c b/src/backend/postmaster/walsummarizer.c
index 2d8f57099fd..9f8ef8159d1 100644
--- a/src/backend/postmaster/walsummarizer.c
+++ b/src/backend/postmaster/walsummarizer.c
@@ -48,6 +48,8 @@
#include "storage/shmem.h"
#include "utils/guc.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/wait_event.h"
/*
@@ -249,7 +251,7 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -271,6 +273,11 @@ WalSummarizerMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* If an exception is encountered, processing resumes here.
*/
diff --git a/src/backend/postmaster/walwriter.c b/src/backend/postmaster/walwriter.c
index 23e79a32345..ded0f250288 100644
--- a/src/backend/postmaster/walwriter.c
+++ b/src/backend/postmaster/walwriter.c
@@ -61,7 +61,9 @@
#include "storage/smgr.h"
#include "utils/hsearch.h"
#include "utils/memutils.h"
+#include "utils/pgstat_internal.h"
#include "utils/resowner.h"
+#include "utils/timeout.h"
/*
@@ -106,7 +108,7 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SignalHandlerForShutdownRequest);
pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
@@ -116,6 +118,11 @@ WalWriterMain(const void *startup_data, size_t startup_data_len)
*/
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/*
* Create a memory context that we will do all our work in. We do this so
* that we can reset the context during error recovery and thereby avoid
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 10e64a7d1f4..11b7c114d3b 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -77,7 +77,9 @@
#include "utils/builtins.h"
#include "utils/guc.h"
#include "utils/pg_lsn.h"
+#include "utils/pgstat_internal.h"
#include "utils/ps_status.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -252,7 +254,7 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, die); /* request shutdown */
/* SIGQUIT handler was already set up by InitPostmasterChild */
- pqsignal(SIGALRM, SIG_IGN);
+ InitializeTimeouts(); /* establishes SIGALRM handler */
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, procsignal_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
@@ -260,6 +262,11 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len)
/* Reset some signals that are accepted by postmaster but not here */
pqsignal(SIGCHLD, SIG_DFL);
+ /*
+ * Register timeouts needed
+ */
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT, AnytimeStatsUpdateTimeoutHandler);
+
/* Load the libpq-specific functions */
load_file("libpqwalreceiver", false);
if (WalReceiverFunctions == NULL)
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 2cde8ebc729..a7214d0dc6f 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1987,8 +1987,8 @@ WalSndWaitForWal(XLogRecPtr loc)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
@@ -3016,8 +3016,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (TimestampDifferenceExceeds(last_flush, now,
WALSENDER_STATS_FLUSH_INTERVAL))
{
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
last_flush = now;
}
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 21de158adbb..2089de782d5 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3564,6 +3564,18 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..a4ff64dc5ce 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -112,6 +112,7 @@
#include "utils/guc_hooks.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
#include "utils/timestamp.h"
@@ -122,8 +123,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +186,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +288,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +306,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +323,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +339,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +357,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +375,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -436,6 +442,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +460,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +478,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +784,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1293,7 +1290,8 @@ pgstat_prep_pending_entry(PgStat_Kind kind, Oid dboid, uint64 objid, bool *creat
if (entry_ref->pending == NULL)
{
- size_t entrysize = pgstat_get_kind_info(kind)->pending_size;
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+ size_t entrysize = kind_info->pending_size;
Assert(entrysize != (size_t) -1);
@@ -1345,9 +1343,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,8 +1380,22 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
- did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
+ did_flush = kind_info->flush_pending_cb(entry_ref, nowait, anytime_only);
Assert(did_flush || nowait);
@@ -1402,6 +1419,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait, anytime_only);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2163,31 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
+
+/*
+ * Timeout handler for flushing anytime stats.
+ */
+void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index f2f8d3ff75f..b09316d3ab3 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -31,6 +31,7 @@
#include "storage/procarray.h"
#include "utils/memutils.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
/*
* Backend statistics counts waiting to be flushed out. These counters may be
@@ -66,6 +67,9 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context,
INSTR_TIME_ADD(PendingBackendStats.pending_io.pending_times[io_object][io_context][io_op],
io_time);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -82,6 +86,9 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context,
PendingBackendStats.pending_io.counts[io_object][io_context][io_op] += cnt;
PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes;
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
backend_has_iostats = true;
pgstat_report_fixed = true;
}
@@ -268,7 +275,7 @@ pgstat_flush_backend_entry_wal(PgStat_EntryRef *entry_ref)
* if some statistics could not be flushed due to lock contention.
*/
bool
-pgstat_flush_backend(bool nowait, bits32 flags)
+pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only)
{
PgStat_EntryRef *entry_ref;
bool has_pending_data = false;
@@ -311,9 +318,9 @@ pgstat_flush_backend(bool nowait, bits32 flags)
* If some stats could not be flushed due to lock contention, return true.
*/
bool
-pgstat_backend_flush_cb(bool nowait)
+pgstat_backend_flush_cb(bool nowait, bool anytime_only)
{
- return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL);
+ return pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_ALL, anytime_only);
}
/*
diff --git a/src/backend/utils/activity/pgstat_bgwriter.c b/src/backend/utils/activity/pgstat_bgwriter.c
index ed2fd801189..1c5f0c3ec40 100644
--- a/src/backend/utils/activity/pgstat_bgwriter.c
+++ b/src/backend/utils/activity/pgstat_bgwriter.c
@@ -61,7 +61,7 @@ pgstat_report_bgwriter(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_checkpointer.c b/src/backend/utils/activity/pgstat_checkpointer.c
index 1f70194b7a7..2d89a082464 100644
--- a/src/backend/utils/activity/pgstat_checkpointer.c
+++ b/src/backend/utils/activity/pgstat_checkpointer.c
@@ -68,7 +68,7 @@ pgstat_report_checkpointer(void)
/*
* Report IO statistics
*/
- pgstat_flush_io(false);
+ pgstat_flush_io(false, true);
}
/*
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 933dcb5cae5..8e86df60461 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -435,7 +435,7 @@ pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStatShared_Database *sharedent;
PgStat_StatDBEntry *pendingent;
diff --git a/src/backend/utils/activity/pgstat_function.c b/src/backend/utils/activity/pgstat_function.c
index e6b84283c6c..5ba4958382f 100644
--- a/src/backend/utils/activity/pgstat_function.c
+++ b/src/backend/utils/activity/pgstat_function.c
@@ -190,11 +190,13 @@ pgstat_end_function_usage(PgStat_FunctionCallUsage *fcu, bool finalize)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_FunctionCounts *localent;
PgStatShared_Function *shfuncent;
+ Assert(!anytime_only);
+
localent = (PgStat_FunctionCounts *) entry_ref->pending;
shfuncent = (PgStatShared_Function *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 28de24538dc..7cd32900236 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -19,6 +19,7 @@
#include "executor/instrument.h"
#include "storage/bufmgr.h"
#include "utils/pgstat_internal.h"
+#include "utils/timeout.h"
static PgStat_PendingIO PendingIOStats;
static bool have_iostats = false;
@@ -79,6 +80,9 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op,
/* Add the per-backend counts */
pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes);
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_iostats = true;
pgstat_report_fixed = true;
}
@@ -172,9 +176,9 @@ pgstat_fetch_stat_io(void)
* Simpler wrapper of pgstat_io_flush_cb()
*/
void
-pgstat_flush_io(bool nowait)
+pgstat_flush_io(bool nowait, bool anytime_only)
{
- (void) pgstat_io_flush_cb(nowait);
+ (void) pgstat_io_flush_cb(nowait, anytime_only);
}
/*
@@ -186,7 +190,7 @@ pgstat_flush_io(bool nowait)
* acquired. Otherwise, return false.
*/
bool
-pgstat_io_flush_cb(bool nowait)
+pgstat_io_flush_cb(bool nowait, bool anytime_only)
{
LWLock *bktype_lock;
PgStat_BktypeIO *bktype_shstats;
diff --git a/src/backend/utils/activity/pgstat_relation.c b/src/backend/utils/activity/pgstat_relation.c
index bc8c43b96aa..04d21483d93 100644
--- a/src/backend/utils/activity/pgstat_relation.c
+++ b/src/backend/utils/activity/pgstat_relation.c
@@ -267,8 +267,8 @@ pgstat_report_vacuum(Relation rel, PgStat_Counter livetuples,
* is done -- which will likely vacuum many relations -- or until the
* VACUUM command has processed all tables and committed.
*/
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -362,8 +362,8 @@ pgstat_report_analyze(Relation rel,
pgstat_unlock_entry(entry_ref);
/* see pgstat_report_vacuum() */
- pgstat_flush_io(false);
- (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(false, true);
+ (void) pgstat_flush_backend(false, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -812,7 +812,7 @@ pgstat_twophase_postabort(FullTransactionId fxid, uint16 info,
* entry when successfully flushing.
*/
bool
-pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
Oid dboid;
PgStat_TableStatus *lstats; /* pending stats entry */
@@ -820,6 +820,8 @@ pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PgStat_StatTabEntry *tabentry; /* table entry of shared stats */
PgStat_StatDBEntry *dbentry; /* pending database entry */
+ Assert(!anytime_only);
+
dboid = entry_ref->shared_entry->key.dboid;
lstats = (PgStat_TableStatus *) entry_ref->pending;
shtabstats = (PgStatShared_Relation *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c
index 2190f388eae..bf8a4d58673 100644
--- a/src/backend/utils/activity/pgstat_slru.c
+++ b/src/backend/utils/activity/pgstat_slru.c
@@ -19,6 +19,7 @@
#include "utils/pgstat_internal.h"
#include "utils/timestamp.h"
+#include "utils/timeout.h"
static inline PgStat_SLRUStats *get_slru_entry(int slru_idx);
@@ -139,7 +140,7 @@ pgstat_get_slru_index(const char *name)
* acquired. Otherwise return false.
*/
bool
-pgstat_slru_flush_cb(bool nowait)
+pgstat_slru_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_SLRU *stats_shmem = &pgStatLocal.shmem->slru;
int i;
@@ -223,6 +224,9 @@ get_slru_entry(int slru_idx)
Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS));
+ /* Schedule next anytime stats update timeout */
+ pgstat_schedule_anytime_update();
+
have_slrustats = true;
pgstat_report_fixed = true;
diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c
index 500b1899188..c4614817966 100644
--- a/src/backend/utils/activity/pgstat_subscription.c
+++ b/src/backend/utils/activity/pgstat_subscription.c
@@ -116,11 +116,13 @@ pgstat_fetch_stat_subscription(Oid subid)
* false without flushing the entry. Otherwise returns true.
*/
bool
-pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
+pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_BackendSubEntry *localent;
PgStatShared_Subscription *shsubent;
+ Assert(!anytime_only);
+
localent = (PgStat_BackendSubEntry *) entry_ref->pending;
shsubent = (PgStatShared_Subscription *) entry_ref->shared_stats;
diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c
index 183e0a7a97b..2c2f3f10e10 100644
--- a/src/backend/utils/activity/pgstat_wal.c
+++ b/src/backend/utils/activity/pgstat_wal.c
@@ -51,12 +51,12 @@ pgstat_report_wal(bool force)
nowait = !force;
/* flush wal stats */
- (void) pgstat_wal_flush_cb(nowait);
- pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL);
+ (void) pgstat_wal_flush_cb(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_WAL, true);
/* flush IO stats */
- pgstat_flush_io(nowait);
- (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO);
+ pgstat_flush_io(nowait, true);
+ (void) pgstat_flush_backend(nowait, PGSTAT_BACKEND_FLUSH_IO, true);
}
/*
@@ -88,7 +88,7 @@ pgstat_wal_have_pending(void)
* acquired. Otherwise return false.
*/
bool
-pgstat_wal_flush_cb(bool nowait)
+pgstat_wal_flush_cb(bool nowait, bool anytime_only)
{
PgStatShared_Wal *stats_shmem = &pgStatLocal.shmem->wal;
WalUsage wal_usage_diff = {0};
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index b59e08605cc..eeeac1bf39a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -64,6 +64,7 @@
#include "utils/injection_point.h"
#include "utils/memutils.h"
#include "utils/pg_locale.h"
+#include "utils/pgstat_internal.h"
#include "utils/portal.h"
#include "utils/ps_status.h"
#include "utils/snapmgr.h"
@@ -773,6 +774,8 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
}
/*
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index f16f35659b9..84e698da214 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..b340a680614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,8 +536,21 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
+/*
+ * Schedule the next anytime stats update timeout.
+ *
+ * This should be called whenever accumulating statistics that support
+ * FLUSH_ANYTIME flushing mode.
+ */
+#define pgstat_schedule_anytime_update() \
+ do { \
+ if (IsUnderPostmaster && !get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT)) \
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL); \
+ } while (0)
+
extern void pgstat_reset_counters(void);
extern void pgstat_reset(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_reset_of_kind(PgStat_Kind kind);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..607f4255268 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,16 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * The mode of when to flush stats. See PgStat_FlushMode for more details.
+ *
+ * This member only has meaning for statistics kinds that accumulate
+ * pending stats and use flush callbacks. For kinds that write directly to
+ * shared memory (e.g., archiver, bgwriter, checkpointer), this member has
+ * no effect.
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
@@ -297,8 +320,10 @@ typedef struct PgStat_KindInfo
* For variable-numbered stats: flush pending stats. Required if pending
* data is used. See flush_static_cb when dealing with stats data that
* that cannot use PgStat_EntryRef->pending.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait);
+ bool (*flush_pending_cb) (PgStat_EntryRef *sr, bool nowait, bool anytime_only);
/*
* For variable-numbered stats: delete pending stats. Optional.
@@ -366,8 +391,10 @@ typedef struct PgStat_KindInfo
*
* "pgstat_report_fixed" needs to be set to trigger the flush of pending
* stats.
+ *
+ * The anytime_only parameter indicates whether this is an anytime flush.
*/
- bool (*flush_static_cb) (bool nowait);
+ bool (*flush_static_cb) (bool nowait, bool anytime_only);
/*
* For fixed-numbered statistics: Reset All.
@@ -677,6 +704,7 @@ extern PgStat_EntryRef *pgstat_fetch_pending_entry(PgStat_Kind kind,
extern void *pgstat_fetch_entry(PgStat_Kind kind, Oid dboid, uint64 objid);
extern void pgstat_snapshot_fixed(PgStat_Kind kind);
+extern void AnytimeStatsUpdateTimeoutHandler(void);
/*
@@ -696,8 +724,8 @@ extern void pgstat_archiver_snapshot_cb(void);
#define PGSTAT_BACKEND_FLUSH_WAL (1 << 1) /* Flush WAL statistics */
#define PGSTAT_BACKEND_FLUSH_ALL (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL)
-extern bool pgstat_flush_backend(bool nowait, bits32 flags);
-extern bool pgstat_backend_flush_cb(bool nowait);
+extern bool pgstat_flush_backend(bool nowait, bits32 flags, bool anytime_only);
+extern bool pgstat_backend_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header,
TimestampTz ts);
@@ -729,7 +757,7 @@ extern void AtEOXact_PgStat_Database(bool isCommit, bool parallel);
extern PgStat_StatDBEntry *pgstat_prep_database_pending(Oid dboid);
extern void pgstat_reset_database_timestamp(Oid dboid, TimestampTz ts);
-extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -737,7 +765,7 @@ extern void pgstat_database_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_function.c
*/
-extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -745,9 +773,9 @@ extern void pgstat_function_reset_timestamp_cb(PgStatShared_Common *header, Time
* Functions in pgstat_io.c
*/
-extern void pgstat_flush_io(bool nowait);
+extern void pgstat_flush_io(bool nowait, bool anytime_only);
-extern bool pgstat_io_flush_cb(bool nowait);
+extern bool pgstat_io_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_io_init_shmem_cb(void *stats);
extern void pgstat_io_reset_all_cb(TimestampTz ts);
extern void pgstat_io_snapshot_cb(void);
@@ -762,7 +790,7 @@ extern void AtEOSubXact_PgStat_Relations(PgStat_SubXactStatus *xact_state, bool
extern void AtPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
extern void PostPrepare_PgStat_Relations(PgStat_SubXactStatus *xact_state);
-extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_relation_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_relation_delete_pending_cb(PgStat_EntryRef *entry_ref);
extern void pgstat_relation_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
@@ -809,7 +837,7 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind,
* Functions in pgstat_slru.c
*/
-extern bool pgstat_slru_flush_cb(bool nowait);
+extern bool pgstat_slru_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_slru_init_shmem_cb(void *stats);
extern void pgstat_slru_reset_all_cb(TimestampTz ts);
extern void pgstat_slru_snapshot_cb(void);
@@ -820,7 +848,7 @@ extern void pgstat_slru_snapshot_cb(void);
*/
extern void pgstat_wal_init_backend_cb(void);
-extern bool pgstat_wal_flush_cb(bool nowait);
+extern bool pgstat_wal_flush_cb(bool nowait, bool anytime_only);
extern void pgstat_wal_init_shmem_cb(void *stats);
extern void pgstat_wal_reset_all_cb(TimestampTz ts);
extern void pgstat_wal_snapshot_cb(void);
@@ -830,7 +858,7 @@ extern void pgstat_wal_snapshot_cb(void);
* Functions in pgstat_subscription.c
*/
-extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait);
+extern bool pgstat_subscription_flush_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only);
extern void pgstat_subscription_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts);
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/test/modules/test_custom_stats/test_custom_var_stats.c b/src/test/modules/test_custom_stats/test_custom_var_stats.c
index 64a8fe63cce..bc0b5d6e0eb 100644
--- a/src/test/modules/test_custom_stats/test_custom_var_stats.c
+++ b/src/test/modules/test_custom_stats/test_custom_var_stats.c
@@ -83,7 +83,7 @@ static dsa_area *custom_stats_description_dsa = NULL;
/* Flush callback: merge pending stats into shared memory */
static bool test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref,
- bool nowait);
+ bool nowait, bool anytime_only);
/* Serialization callback: write auxiliary entry data */
static void test_custom_stats_var_to_serialized_data(const PgStat_HashKey *key,
@@ -150,7 +150,7 @@ _PG_init(void)
* Returns false only if nowait=true and lock acquisition fails.
*/
static bool
-test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait)
+test_custom_stats_var_flush_pending_cb(PgStat_EntryRef *entry_ref, bool nowait, bool anytime_only)
{
PgStat_StatCustomVarEntry *pending_entry;
PgStatShared_CustomVarEntry *shared_entry;
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 241945734ec..1dbc4b96f51 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2271,6 +2271,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--C2xzVbxFmVrP7k6O
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v7-0002-Add-anytime-flush-tests-for-custom-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v4 1/4] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- This relies on the existing PGSTAT_MIN_INTERVAL to fire every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/storage/lmgr/proc.c | 10 +++
src/backend/tcop/postgres.c | 16 ++++
src/backend/utils/activity/pgstat.c | 111 +++++++++++++++++++++++-----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 20 +++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
10 files changed, 162 insertions(+), 18 deletions(-)
7.0% src/backend/storage/lmgr/
7.3% src/backend/tcop/
61.1% src/backend/utils/activity/
8.6% src/backend/utils/init/
11.8% src/include/utils/
3.6% src/include/
diff --git a/src/backend/storage/lmgr/proc.c b/src/backend/storage/lmgr/proc.c
index 063826ae576..012705a2ee6 100644
--- a/src/backend/storage/lmgr/proc.c
+++ b/src/backend/storage/lmgr/proc.c
@@ -1322,6 +1322,7 @@ ProcSleep(LOCALLOCK *locallock)
bool allow_autovacuum_cancel = true;
bool logged_recovery_conflict = false;
ProcWaitStatus myWaitStatus;
+ bool anytime_timeout_was_active = false;
/* The caller must've armed the on-error cleanup mechanism */
Assert(GetAwaitedLock() == locallock);
@@ -1398,6 +1399,12 @@ ProcSleep(LOCALLOCK *locallock)
standbyWaitStart = GetCurrentTimestamp();
}
+ anytime_timeout_was_active = get_timeout_active(ANYTIME_STATS_UPDATE_TIMEOUT);
+
+ /* No need to try to flush the statistics while the process is sleeping */
+ if (anytime_timeout_was_active)
+ disable_timeout(ANYTIME_STATS_UPDATE_TIMEOUT, false);
+
/*
* If somebody wakes us between LWLockRelease and WaitLatch, the latch
* will not wait. But a set latch does not necessarily mean that the lock
@@ -1661,6 +1668,9 @@ ProcSleep(LOCALLOCK *locallock)
}
} while (myWaitStatus == PROC_WAIT_STATUS_WAITING);
+ if (anytime_timeout_was_active)
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
+
/*
* Disable the timers, if they are still running. As in LockErrorCleanup,
* we must preserve the LOCK_TIMEOUT indicator flag: if a lock timeout has
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..132fae61423 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,22 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_MIN_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ab4d9088a9a 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -122,8 +122,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +185,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +287,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +305,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +322,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +338,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +356,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +374,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +393,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +410,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +427,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +444,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +462,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +480,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +786,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1344,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1381,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1402,6 +1420,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2164,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..6076f531c4a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..1651f16f966 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..9ca39ea9a9a 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,13 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /*
+ * Some stats have to be updated only at transaction boundaries (such as
+ * tuples_inserted updated, deleted), so it's very important to set the
+ * right flush mode (FLUSH_AT_TXN_BOUNDARY being the default).
+ */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ddbe4c64971..af21c87234a 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--wR/mWGXukst4NraF
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v4-0002-Add-GUC-to-specify-non-transactional-statistics-f.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v3 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushMode enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- This relies on the existing PGSTAT_MIN_INTERVAL to fire every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/tcop/postgres.c | 16 ++++
src/backend/utils/activity/pgstat.c | 111 +++++++++++++++++++++++-----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 16 ++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
9 files changed, 148 insertions(+), 18 deletions(-)
8.1% src/backend/tcop/
68.4% src/backend/utils/activity/
9.6% src/backend/utils/init/
9.2% src/include/utils/
4.1% src/include/
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..132fae61423 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,22 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_MIN_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..ab4d9088a9a 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -122,8 +122,6 @@
* ----------
*/
-/* minimum interval non-forced stats flushes.*/
-#define PGSTAT_MIN_INTERVAL 1000
/* how long until to block flushing pending stats updates */
#define PGSTAT_MAX_INTERVAL 60000
/* when to call pgstat_report_stat() again, even when idle */
@@ -187,7 +185,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +287,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +305,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +322,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +338,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +356,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_mode = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +374,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_mode = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +393,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +410,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +427,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +444,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +462,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +480,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_mode = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +786,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1344,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1381,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1402,6 +1420,33 @@ pgstat_flush_pending_entries(bool nowait)
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_mode == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2164,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..6076f531c4a 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_MIN_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..1651f16f966 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* Minimum interval non-forced stats flushes */
+#define PGSTAT_MIN_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..46ce90c9624 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,19 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush mode for statistics kinds.
+ *
+ * FLUSH_AT_TXN_BOUNDARY has to be the first because we want it to be the
+ * default value.
+ */
+typedef enum PgStat_FlushMode
+{
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+} PgStat_FlushMode;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +264,9 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /* Flush mode */
+ PgStat_FlushMode flush_mode;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 3f3a888fd0e..d3912b43fdc 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushMode
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--lCc1MWfC7i2uqs8g
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v3-0002-Remove-useless-calls-to-flush-some-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
* [PATCH v2 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing
@ 2026-01-05 09:41 Bertrand Drouvot <[email protected]>
0 siblings, 0 replies; 166+ messages in thread
From: Bertrand Drouvot @ 2026-01-05 09:41 UTC (permalink / raw)
Long running transactions can accumulate significant statistics (WAL, IO, ...)
that remain unflushed until the transaction ends. This delays visibility of
resource usage in monitoring views like pg_stat_io and pg_stat_wal.
This commit introduces pgstat_report_anytime_stat(), which flushes
non transactional statistics even inside active transactions. A new timeout
handler fires every second to call this function, ensuring timely stats visibility
without waiting for transaction completion.
Implementation details:
- Add PgStat_FlushBehavior enum to classify stats kinds:
* FLUSH_ANYTIME: Stats that can always be flushed (WAL, IO, ...)
* FLUSH_AT_TXN_BOUNDARY: Stats requiring transaction boundaries
- Modify pgstat_flush_pending_entries() and pgstat_flush_fixed_stats()
to accept a boolean anytime_only parameter:
* When false: flushes all stats (existing behavior)
* When true: flushes only FLUSH_ANYTIME stats and skips FLUSH_AT_TXN_BOUNDARY stats
- Register ANYTIME_STATS_UPDATE_TIMEOUT that fires every 1 second, calling
pgstat_report_anytime_stat(false)
The force parameter in pgstat_report_anytime_stat() is currently unused (always
called with force=false) but reserved for future use cases requiring immediate
flushing.
---
src/backend/tcop/postgres.c | 16 ++++
src/backend/utils/activity/pgstat.c | 113 ++++++++++++++++++++++++----
src/backend/utils/init/globals.c | 1 +
src/backend/utils/init/postinit.c | 15 ++++
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 4 +
src/include/utils/pgstat_internal.h | 13 ++++
src/include/utils/timeout.h | 1 +
src/tools/pgindent/typedefs.list | 1 +
9 files changed, 149 insertions(+), 16 deletions(-)
8.2% src/backend/tcop/
69.3% src/backend/utils/activity/
9.7% src/backend/utils/init/
7.7% src/include/utils/
4.5% src/include/
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index e54bf1e760f..9c4a9078ee0 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -3530,6 +3530,22 @@ ProcessInterrupts(void)
pgstat_report_stat(true);
}
+ /*
+ * Flush stats outside of transaction boundary if the timeout fired.
+ * Unlike transactional stats, these can be flushed even inside a running
+ * transaction.
+ */
+ if (AnytimeStatsUpdateTimeoutPending)
+ {
+ AnytimeStatsUpdateTimeoutPending = false;
+
+ pgstat_report_anytime_stat(false);
+
+ /* Schedule next timeout */
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT,
+ PGSTAT_ANYTIME_FLUSH_INTERVAL);
+ }
+
if (ProcSignalBarrierPending)
ProcessProcSignalBarrier();
diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c
index 11bb71cad5a..0f45a7d165e 100644
--- a/src/backend/utils/activity/pgstat.c
+++ b/src/backend/utils/activity/pgstat.c
@@ -187,7 +187,8 @@ static void pgstat_init_snapshot_fixed(void);
static void pgstat_reset_after_failure(void);
-static bool pgstat_flush_pending_entries(bool nowait);
+static bool pgstat_flush_pending_entries(bool nowait, bool anytime_only);
+static bool pgstat_flush_fixed_stats(bool nowait, bool anytime_only);
static void pgstat_prep_snapshot(void);
static void pgstat_build_snapshot(void);
@@ -288,6 +289,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_database entries can be seen in all databases */
.accessed_across_databases = true,
@@ -305,6 +307,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Relation),
.shared_data_off = offsetof(PgStatShared_Relation, stats),
@@ -321,6 +324,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.shared_size = sizeof(PgStatShared_Function),
.shared_data_off = offsetof(PgStatShared_Function, stats),
@@ -336,6 +340,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
.accessed_across_databases = true,
@@ -353,6 +358,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = true,
+ .flush_behavior = FLUSH_AT_TXN_BOUNDARY,
/* so pg_stat_subscription_stats entries can be seen in all databases */
.accessed_across_databases = true,
@@ -370,6 +376,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = false,
.write_to_file = false,
+ .flush_behavior = FLUSH_ANYTIME,
.accessed_across_databases = true,
@@ -388,6 +395,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, archiver),
.shared_ctl_off = offsetof(PgStat_ShmemControl, archiver),
@@ -404,6 +412,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, bgwriter),
.shared_ctl_off = offsetof(PgStat_ShmemControl, bgwriter),
@@ -420,6 +429,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, checkpointer),
.shared_ctl_off = offsetof(PgStat_ShmemControl, checkpointer),
@@ -436,6 +446,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, io),
.shared_ctl_off = offsetof(PgStat_ShmemControl, io),
@@ -453,6 +464,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, slru),
.shared_ctl_off = offsetof(PgStat_ShmemControl, slru),
@@ -470,6 +482,7 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE]
.fixed_amount = true,
.write_to_file = true,
+ .flush_behavior = FLUSH_ANYTIME,
.snapshot_ctl_off = offsetof(PgStat_Snapshot, wal),
.shared_ctl_off = offsetof(PgStat_ShmemControl, wal),
@@ -775,23 +788,11 @@ pgstat_report_stat(bool force)
partial_flush = false;
/* flush of variable-numbered stats tracked in pending entries list */
- partial_flush |= pgstat_flush_pending_entries(nowait);
+ partial_flush |= pgstat_flush_pending_entries(nowait, false);
/* flush of other stats kinds */
if (pgstat_report_fixed)
- {
- for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
- {
- const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
-
- if (!kind_info)
- continue;
- if (!kind_info->flush_static_cb)
- continue;
-
- partial_flush |= kind_info->flush_static_cb(nowait);
- }
- }
+ partial_flush |= pgstat_flush_fixed_stats(nowait, false);
last_flush = now;
@@ -1345,9 +1346,14 @@ pgstat_delete_pending_entry(PgStat_EntryRef *entry_ref)
/*
* Flush out pending variable-numbered stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME entries.
+ * This is safe to call inside transactions.
+ *
+ * If anytime_only is false, flushes all entries.
*/
static bool
-pgstat_flush_pending_entries(bool nowait)
+pgstat_flush_pending_entries(bool nowait, bool anytime_only)
{
bool have_pending = false;
dlist_node *cur = NULL;
@@ -1377,6 +1383,20 @@ pgstat_flush_pending_entries(bool nowait)
Assert(!kind_info->fixed_amount);
Assert(kind_info->flush_pending_cb != NULL);
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ {
+ have_pending = true;
+
+ if (dlist_has_next(&pgStatPending, cur))
+ next = dlist_next_node(&pgStatPending, cur);
+ else
+ next = NULL;
+
+ cur = next;
+ continue;
+ }
+
/* flush the stats, if possible */
did_flush = kind_info->flush_pending_cb(entry_ref, nowait);
@@ -1397,11 +1417,42 @@ pgstat_flush_pending_entries(bool nowait)
cur = next;
}
+ /*
+ * When in anytime_only mode, the list may not be empty because
+ * FLUSH_AT_TXN_BOUNDARY entries were skipped.
+ */
Assert(dlist_is_empty(&pgStatPending) == !have_pending);
return have_pending;
}
+/*
+ * Flush fixed-amount stats.
+ *
+ * If anytime_only is true, only flushes FLUSH_ANYTIME stats (safe inside transactions).
+ * If anytime_only is false, flushes all stats with flush_static_cb.
+ */
+static bool
+pgstat_flush_fixed_stats(bool nowait, bool anytime_only)
+{
+ bool partial_flush = false;
+
+ for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++)
+ {
+ const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind);
+
+ if (!kind_info || !kind_info->flush_static_cb)
+ continue;
+
+ /* Skip transactional stats if we're in anytime_only mode */
+ if (anytime_only && kind_info->flush_behavior == FLUSH_AT_TXN_BOUNDARY)
+ continue;
+
+ partial_flush |= kind_info->flush_static_cb(nowait);
+ }
+
+ return partial_flush;
+}
/* ------------------------------------------------------------
* Helper / infrastructure functions
@@ -2119,3 +2170,33 @@ assign_stats_fetch_consistency(int newval, void *extra)
if (pgstat_fetch_consistency != newval)
force_stats_snapshot_clear = true;
}
+
+/*
+ * Flushes only FLUSH_ANYTIME stats using non-blocking locks. Transactional
+ * stats (FLUSH_AT_TXN_BOUNDARY) remain pending until transaction boundary.
+ * Safe to call inside transactions.
+ */
+void
+pgstat_report_anytime_stat(bool force)
+{
+ bool nowait = !force;
+
+ pgstat_assert_is_up();
+
+ /*
+ * Exit if no pending stats at all. This avoids unnecessary work when
+ * backends are idle or in sessions without stats accumulation.
+ *
+ * Note: This check isn't precise as there might be only transactional
+ * stats pending, which we'll skip during the flush. However, maintaining
+ * precise tracking would add complexity that does not seem worth it from
+ * a performance point of view (no noticeable performance regression has
+ * been observed with the current implementation).
+ */
+ if (dlist_is_empty(&pgStatPending) && !pgstat_report_fixed)
+ return;
+
+ /* Flush stats outside of transaction boundary */
+ pgstat_flush_pending_entries(nowait, true);
+ pgstat_flush_fixed_stats(nowait, true);
+}
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 36ad708b360..ad44826c39e 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -40,6 +40,7 @@ volatile sig_atomic_t IdleSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
volatile sig_atomic_t LogMemoryContextPending = false;
volatile sig_atomic_t IdleStatsUpdateTimeoutPending = false;
+volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/backend/utils/init/postinit.c b/src/backend/utils/init/postinit.c
index 3f401faf3de..cb0f6aecad1 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -82,6 +82,7 @@ static void TransactionTimeoutHandler(void);
static void IdleSessionTimeoutHandler(void);
static void IdleStatsUpdateTimeoutHandler(void);
static void ClientCheckTimeoutHandler(void);
+static void AnytimeStatsUpdateTimeoutHandler(void);
static bool ThereIsAtLeastOneRole(void);
static void process_startup_options(Port *port, bool am_superuser);
static void process_settings(Oid databaseid, Oid roleid);
@@ -765,6 +766,9 @@ InitPostgres(const char *in_dbname, Oid dboid,
RegisterTimeout(CLIENT_CONNECTION_CHECK_TIMEOUT, ClientCheckTimeoutHandler);
RegisterTimeout(IDLE_STATS_UPDATE_TIMEOUT,
IdleStatsUpdateTimeoutHandler);
+ RegisterTimeout(ANYTIME_STATS_UPDATE_TIMEOUT,
+ AnytimeStatsUpdateTimeoutHandler);
+ enable_timeout_after(ANYTIME_STATS_UPDATE_TIMEOUT, PGSTAT_ANYTIME_FLUSH_INTERVAL);
}
/*
@@ -1446,3 +1450,14 @@ ThereIsAtLeastOneRole(void)
return result;
}
+
+/*
+ * Timeout handler for flushing non-transactional stats.
+ */
+static void
+AnytimeStatsUpdateTimeoutHandler(void)
+{
+ AnytimeStatsUpdateTimeoutPending = true;
+ InterruptPending = true;
+ SetLatch(MyLatch);
+}
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index db559b39c4d..8aeb9628871 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -96,6 +96,7 @@ extern PGDLLIMPORT volatile sig_atomic_t IdleSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
extern PGDLLIMPORT volatile sig_atomic_t LogMemoryContextPending;
extern PGDLLIMPORT volatile sig_atomic_t IdleStatsUpdateTimeoutPending;
+extern PGDLLIMPORT volatile sig_atomic_t AnytimeStatsUpdateTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t CheckClientConnectionPending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index fff7ecc2533..86e65397614 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -35,6 +35,9 @@
/* Default directory to store temporary statistics data in */
#define PG_STAT_TMP_DIR "pg_stat_tmp"
+/* When to call pgstat_report_anytime_stat() again */
+#define PGSTAT_ANYTIME_FLUSH_INTERVAL 1000
+
/* Values for track_functions GUC variable --- order is significant! */
typedef enum TrackFunctionsLevel
{
@@ -533,6 +536,7 @@ extern void pgstat_initialize(void);
/* Functions called from backends */
extern long pgstat_report_stat(bool force);
+extern void pgstat_report_anytime_stat(bool force);
extern void pgstat_force_next_flush(void);
extern void pgstat_reset_counters(void);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index 9b8fbae00ed..63feae640d1 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -224,6 +224,16 @@ typedef struct PgStat_SubXactStatus
PgStat_TableXactStatus *first; /* head of list for this subxact */
} PgStat_SubXactStatus;
+/*
+ * Flush behavior for statistics kinds.
+ */
+typedef enum PgStat_FlushBehavior
+{
+ FLUSH_ANYTIME, /* All fields can be flushed anytime,
+ * including within transactions */
+ FLUSH_AT_TXN_BOUNDARY, /* All fields can only be flushed at
+ * transaction boundary */
+} PgStat_FlushBehavior;
/*
* Metadata for a specific kind of statistics.
@@ -251,6 +261,9 @@ typedef struct PgStat_KindInfo
*/
bool track_entry_count:1;
+ /* Flush behavior */
+ PgStat_FlushBehavior flush_behavior;
+
/*
* The size of an entry in the shared stats hash table (pointed to by
* PgStatShared_HashEntry->body). For fixed-numbered statistics, this is
diff --git a/src/include/utils/timeout.h b/src/include/utils/timeout.h
index 0965b590b34..10723bb664c 100644
--- a/src/include/utils/timeout.h
+++ b/src/include/utils/timeout.h
@@ -35,6 +35,7 @@ typedef enum TimeoutId
IDLE_SESSION_TIMEOUT,
IDLE_STATS_UPDATE_TIMEOUT,
CLIENT_CONNECTION_CHECK_TIMEOUT,
+ ANYTIME_STATS_UPDATE_TIMEOUT,
STARTUP_PROGRESS_TIMEOUT,
/* First user-definable timeout reason */
USER_TIMEOUT,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 3f3a888fd0e..610b35a9b31 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2268,6 +2268,7 @@ PgStat_Counter
PgStat_EntryRef
PgStat_EntryRefHashEntry
PgStat_FetchConsistency
+PgStat_FlushBehavior
PgStat_FunctionCallUsage
PgStat_FunctionCounts
PgStat_HashKey
--
2.34.1
--OFsrIl+bjhifp5hk
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
filename="v2-0002-Remove-useless-calls-to-flush-some-stats.patch"
^ permalink raw reply [nested|flat] 166+ messages in thread
end of thread, other threads:[~2026-01-05 09:41 UTC | newest]
Thread overview: 166+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2019-07-01 15:11 [PATCH v1] Fix usage of unified logging pg_log_* in pg_rewind and initdb Alexey Kondratov <[email protected]>
2019-11-11 13:49 [PATCH 1/2] Unique key jesperpedersen <[email protected]>
2023-06-25 11:48 [PATCH v1 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-25 11:48 [PATCH v1 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-25 11:48 [PATCH v1 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-25 11:48 [PATCH v1 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-25 11:48 [PATCH v1 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-26 08:05 [PATCH v2 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-26 08:05 [PATCH v2 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-26 08:05 [PATCH v2 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-26 08:05 [PATCH v2 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-06-26 08:05 [PATCH v2 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-07-26 10:49 [PATCH v3 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-07-26 10:49 [PATCH v3 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-07-26 10:49 [PATCH v3 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-07-26 10:49 [PATCH v3 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-08-09 07:56 [PATCH v4 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-08-09 07:56 [PATCH v4 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-08-09 07:56 [PATCH v4 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-08-09 07:56 [PATCH v4 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-08-09 07:56 [PATCH v4 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-08-09 07:56 [PATCH v4 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-02 06:32 [PATCH v5 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-12 05:22 [PATCH v6 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-22 04:53 [PATCH v7 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-22 04:53 [PATCH v7 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-22 04:53 [PATCH v7 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-22 04:53 [PATCH v7 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-22 04:53 [PATCH v7 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-25 05:01 [PATCH v8 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-25 05:01 [PATCH v8 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-25 05:01 [PATCH v8 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-09-25 05:01 [PATCH v8 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-04 05:51 [PATCH v9 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-04 05:51 [PATCH v9 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-04 05:51 [PATCH v9 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-04 05:51 [PATCH v9 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-22 02:22 [PATCH v10 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-22 02:22 [PATCH v10 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-22 02:22 [PATCH v10 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-10-22 02:22 [PATCH v10 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-11-08 06:57 [PATCH v11 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-11-08 06:57 [PATCH v11 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-11-08 06:57 [PATCH v11 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-12-04 11:23 [PATCH v12 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-12-04 11:23 [PATCH v12 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2023-12-04 11:23 [PATCH v12 2/7] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-01-22 09:45 [PATCH v13 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-01-22 09:45 [PATCH v13 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-01-22 09:45 [PATCH v13 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-02-28 13:59 [PATCH v14 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-02-28 13:59 [PATCH v14 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-02-28 13:59 [PATCH v14 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-03-28 10:30 [PATCH v15 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-03-28 10:30 [PATCH v15 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-03-28 10:30 [PATCH v15 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-03-28 10:30 [PATCH v15 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-03-28 10:30 [PATCH v15 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-04-12 06:49 [PATCH v16 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-04-12 06:49 [PATCH v16 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-04-12 06:49 [PATCH v16 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-04-28 11:00 [PATCH v17 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-04-28 11:00 [PATCH v17 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-04-28 11:00 [PATCH v17 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-11 07:11 [PATCH v18 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-11 07:11 [PATCH v18 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-11 07:11 [PATCH v18 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-14 23:26 [PATCH v19 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-14 23:26 [PATCH v19 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-14 23:26 [PATCH v19 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-24 02:26 [PATCH v20 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-24 02:26 [PATCH v20 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-05-24 02:26 [PATCH v20 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-08-26 04:32 [PATCH v21 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-08-26 04:32 [PATCH v21 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-08-26 04:32 [PATCH v21 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-09-19 04:48 [PATCH v22 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-09-19 04:48 [PATCH v22 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-09-19 04:48 [PATCH v22 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-10-02 14:46 Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-03 06:22 ` Re: general purpose array_sort Amit Langote <[email protected]>
2024-10-09 14:10 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-09 15:46 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-10 02:30 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-11 05:12 ` Re: general purpose array_sort jian he <[email protected]>
2024-10-13 14:43 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-13 20:10 ` Re: general purpose array_sort Tom Lane <[email protected]>
2024-10-23 14:28 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-24 12:39 ` Re: general purpose array_sort jian he <[email protected]>
2024-10-24 14:33 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-24 14:57 ` Re: general purpose array_sort Aleksander Alekseev <[email protected]>
2024-10-24 15:11 ` Re: general purpose array_sort David G. Johnston <[email protected]>
2024-10-24 15:27 ` Re: general purpose array_sort Aleksander Alekseev <[email protected]>
2024-10-24 15:40 ` Re: general purpose array_sort David G. Johnston <[email protected]>
2024-10-24 16:06 ` Re: general purpose array_sort Tom Lane <[email protected]>
2024-10-24 16:47 ` Re: general purpose array_sort David G. Johnston <[email protected]>
2024-10-24 17:18 ` Re: general purpose array_sort Aleksander Alekseev <[email protected]>
2024-10-25 12:02 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-26 04:20 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-28 16:48 ` Re: general purpose array_sort Aleksander Alekseev <[email protected]>
2024-10-29 08:28 ` Re: general purpose array_sort jian he <[email protected]>
2024-10-29 12:14 ` Re: general purpose array_sort Aleksander Alekseev <[email protected]>
2024-10-29 14:49 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-30 13:29 ` Re: general purpose array_sort Aleksander Alekseev <[email protected]>
2024-10-30 14:17 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-30 14:41 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-03 03:33 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-04 05:46 ` Re: general purpose array_sort Michael Paquier <[email protected]>
2024-11-04 07:16 ` Re: general purpose array_sort jian he <[email protected]>
2024-11-05 01:12 ` Re: general purpose array_sort Michael Paquier <[email protected]>
2024-11-07 14:06 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-08 00:51 ` Re: general purpose array_sort Michael Paquier <[email protected]>
2024-11-09 14:45 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-07 13:56 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-07 14:29 ` Re: general purpose array_sort Robert Haas <[email protected]>
2024-11-07 14:51 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-08 00:01 ` Re: general purpose array_sort Michael Paquier <[email protected]>
2024-11-04 11:33 ` Re: general purpose array_sort Dean Rasheed <[email protected]>
2024-11-05 07:12 ` Re: general purpose array_sort jian he <[email protected]>
2024-11-05 12:30 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-11-06 07:10 ` Re: general purpose array_sort jian he <[email protected]>
2024-11-07 14:26 ` Re: general purpose array_sort Junwang Zhao <[email protected]>
2024-10-25 03:56 [PATCH v23 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-10-25 03:56 [PATCH v23 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-10-25 03:56 [PATCH v23 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-19 06:06 [PATCH v24 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-19 06:06 [PATCH v24 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-19 06:06 [PATCH v24 2/8] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-21 06:19 [PATCH v25 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-21 06:19 [PATCH v25 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-21 06:19 [PATCH v25 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-30 12:44 [PATCH v26 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-30 12:44 [PATCH v26 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-30 12:44 [PATCH v26 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-30 23:53 [PATCH v27 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-30 23:53 [PATCH v27 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2024-12-30 23:53 [PATCH v27 2/9] Row pattern recognition patch (parse/analysis). Tatsuo Ishii <[email protected]>
2026-01-05 09:41 [PATCH v8 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v4 1/4] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v11 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v11 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v5 1/4] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v1 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v7 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v1 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v10 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v9 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v2 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v6 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v3 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v4 1/4] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v2 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v3 1/3] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
2026-01-05 09:41 [PATCH v7 1/5] Add pgstat_report_anytime_stat() for periodic stats flushing Bertrand Drouvot <[email protected]>
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox