agora inbox for [email protected]  
help / color / mirror / Atom feed
[PATCH v28 04/11] pg_ls_tmpdir to show directories and "isdir" argument..
4+ messages / 2 participants
[nested] [flat]

* [PATCH v28 04/11] pg_ls_tmpdir to show directories and "isdir" argument..
@ 2020-03-09 03:57 Justin Pryzby <[email protected]>
  0 siblings, 0 replies; 4+ messages in thread

From: Justin Pryzby @ 2020-03-09 03:57 UTC (permalink / raw)

similar to pg_stat_file().

It's worth breaking the function's return type, since core postgres creates
"shared filesets" underneath the temp dirs, and it's unreasonable to not show
them here, and the alternative query to show them is unreasaonably complicated.

See following commit which also adds these columns to the other pg_ls_*
functions.  Although I don't think it matters that they're easily UNIONed, it'd
still make great sense if they returned the same columns.

Need catversion bump
---
 doc/src/sgml/func.sgml                       | 17 +++++++++--------
 src/backend/utils/adt/genfile.c              |  2 +-
 src/include/catalog/pg_proc.dat              |  8 ++++----
 src/test/regress/expected/misc_functions.out |  4 ++--
 src/test/regress/output/tablespace.source    |  4 ++--
 5 files changed, 18 insertions(+), 17 deletions(-)

diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index 507a6d73f8..940c653807 100644
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -26950,16 +26950,17 @@ SELECT pg_size_pretty(sum(pg_relation_size(relid))) AS total_size
         <returnvalue>setof record</returnvalue>
         ( <parameter>name</parameter> <type>text</type>,
         <parameter>size</parameter> <type>bigint</type>,
-        <parameter>modification</parameter> <type>timestamp with time zone</type> )
+        <parameter>modification</parameter> <type>timestamp with time zone</type>,
+        <parameter>isdir</parameter> <type>boolean</type> )
        </para>
        <para>
-        Returns the name, size, and last modification time (mtime) of each
-        ordinary file in the temporary file directory for the
-        specified <parameter>tablespace</parameter>.
-        If <parameter>tablespace</parameter> is not provided,
-        the <literal>pg_default</literal> tablespace is examined.  Filenames
-        beginning with a dot, directories, and other special files are
-        excluded.
+        For each file in the temporary directory within the given
+        <parameter>tablespace</parameter>, return the file's name, size, last
+        modification time (mtime) and a boolean indicating if the file is a directory.
+        Directories are used for temporary files shared by parallel processes.
+        If <parameter>tablespace</parameter> is not provided, the
+        <literal>pg_default</literal> tablespace is examined.
+        Filenames beginning with a dot and special file types are excluded.
        </para>
        <para>
         This function is restricted to superusers and members of
diff --git a/src/backend/utils/adt/genfile.c b/src/backend/utils/adt/genfile.c
index f59f853983..a657d8d431 100644
--- a/src/backend/utils/adt/genfile.c
+++ b/src/backend/utils/adt/genfile.c
@@ -709,7 +709,7 @@ pg_ls_tmpdir(FunctionCallInfo fcinfo, Oid tblspc)
 
 	TempTablespacePath(path, tblspc);
 	return pg_ls_dir_files(fcinfo, path,
-			LS_DIR_HISTORIC | LS_DIR_MISSING_OK);
+			LS_DIR_SKIP_HIDDEN | LS_DIR_SKIP_SPECIAL | LS_DIR_ISDIR | LS_DIR_METADATA | LS_DIR_MISSING_OK);
 }
 
 /*
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 550fbf734a..441cbb2cde 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11526,13 +11526,13 @@
 { oid => '5029', descr => 'list files in the pgsql_tmp directory',
   proname => 'pg_ls_tmpdir', procost => '10', prorows => '20', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
-  proallargtypes => '{text,int8,timestamptz}', proargmodes => '{o,o,o}',
-  proargnames => '{name,size,modification}', prosrc => 'pg_ls_tmpdir_noargs' },
+  proallargtypes => '{text,int8,timestamptz,bool}', proargmodes => '{o,o,o,o}',
+  proargnames => '{name,size,modification,isdir}', prosrc => 'pg_ls_tmpdir_noargs' },
 { oid => '5030', descr => 'list files in the pgsql_tmp directory',
   proname => 'pg_ls_tmpdir', procost => '10', prorows => '20', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => 'oid',
-  proallargtypes => '{oid,text,int8,timestamptz}', proargmodes => '{i,o,o,o}',
-  proargnames => '{tablespace,name,size,modification}',
+  proallargtypes => '{oid,text,int8,timestamptz,bool}', proargmodes => '{i,o,o,o,o}',
+  proargnames => '{tablespace,name,size,modification,isdir}',
   prosrc => 'pg_ls_tmpdir_1arg' },
 { oid => '8450', descr => 'list directory with metadata',
   proname => 'pg_ls_dir_metadata', procost => '10', prorows => '20', proretset => 't',
diff --git a/src/test/regress/expected/misc_functions.out b/src/test/regress/expected/misc_functions.out
index b4112ec298..0d3e88fccf 100644
--- a/src/test/regress/expected/misc_functions.out
+++ b/src/test/regress/expected/misc_functions.out
@@ -241,8 +241,8 @@ select * from pg_stat_file('.') limit 0;
 -- This tests the missing_ok parameter, which causes pg_ls_tmpdir to succeed even if the tmpdir doesn't exist yet
 -- The name='' condition is never true, so the function runs to completion but returns zero rows.
 select * from pg_ls_tmpdir() where name='Does not exist';
- name | size | modification 
-------+------+--------------
+ name | size | modification | isdir 
+------+------+--------------+-------
 (0 rows)
 
 select filename, isdir from pg_ls_dir_metadata('.') where filename='.';
diff --git a/src/test/regress/output/tablespace.source b/src/test/regress/output/tablespace.source
index 5f7ce3d09f..4183a77b23 100644
--- a/src/test/regress/output/tablespace.source
+++ b/src/test/regress/output/tablespace.source
@@ -17,8 +17,8 @@ CREATE TABLESPACE regress_tblspace LOCATION '@testtablespace@';
 -- The name='' condition is never true, so the function runs to completion but returns zero rows.
 -- The query is written to ERROR if the tablespace doesn't exist, rather than silently failing to call pg_ls_tmpdir()
 SELECT c.* FROM (SELECT oid FROM pg_tablespace b WHERE b.spcname='regress_tblspace' UNION SELECT 0 ORDER BY 1 DESC LIMIT 1) AS b , pg_ls_tmpdir(oid) AS c WHERE c.name='Does not exist';
- name | size | modification 
-------+------+--------------
+ name | size | modification | isdir 
+------+------+--------------+-------
 (0 rows)
 
 -- try setting and resetting some properties for the new tablespace
-- 
2.17.0


--vk/v8fjDPiDepTtA
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
 filename="v28-0005-pg_ls_-dir-to-show-directories-and-isdir-column.patch"



^ permalink  raw  reply  [nested|flat] 4+ messages in thread

* [PATCH v6 1/3] vacuumdb: Teach vacuum_one_database() to reuse query results.
@ 2025-03-11 16:18 Nathan Bossart <[email protected]>
  0 siblings, 0 replies; 4+ messages in thread

From: Nathan Bossart @ 2025-03-11 16:18 UTC (permalink / raw)

Presently, each call to vacuum_one_database() performs a catalog
query to retrieve the list of tables to process.  A proposed
follow-up commit would add a "missing only" feature to
--analyze-in-stages, which requires us to save the results of the
catalog query (since tables without statistics would have them
after the first stage).  This commit adds this ability via a new
parameter for vacuum_one_database() that specifies either a
previously-retrieved list to process or a place to store the
results of the catalog query for later use.  Note that this commit
does not make use of this new parameter for anything.  That is left
as a future exercise.

The trade-offs of this approach are increased memory usage and less
responsiveness to concurrent catalog changes in later stages,
neither of which is expected to bother anyone.

Author: Corey Huinker <[email protected]>
Co-authored-by: Nathan Bossart <[email protected]>
Reviewed-by: John Naylor <[email protected]>
Discussion: https://postgr.es/m/Z5O1bpcwDrMgyrYy%40nathan
---
 src/bin/scripts/vacuumdb.c | 330 ++++++++++++++++++++++---------------
 1 file changed, 194 insertions(+), 136 deletions(-)

diff --git a/src/bin/scripts/vacuumdb.c b/src/bin/scripts/vacuumdb.c
index 982bf070be6..e28f82a0eba 100644
--- a/src/bin/scripts/vacuumdb.c
+++ b/src/bin/scripts/vacuumdb.c
@@ -62,10 +62,16 @@ typedef enum
 
 static VacObjFilter objfilter = OBJFILTER_NONE;
 
+static SimpleStringList *retrieve_objects(PGconn *conn,
+										  vacuumingOptions *vacopts,
+										  SimpleStringList *objects,
+										  bool echo);
+
 static void vacuum_one_database(ConnParams *cparams,
 								vacuumingOptions *vacopts,
 								int stage,
 								SimpleStringList *objects,
+								SimpleStringList **found_objs,
 								int concurrentCons,
 								const char *progname, bool echo, bool quiet);
 
@@ -405,7 +411,7 @@ main(int argc, char *argv[])
 			{
 				vacuum_one_database(&cparams, &vacopts,
 									stage,
-									&objects,
+									&objects, NULL,
 									concurrentCons,
 									progname, echo, quiet);
 			}
@@ -413,7 +419,7 @@ main(int argc, char *argv[])
 		else
 			vacuum_one_database(&cparams, &vacopts,
 								ANALYZE_NO_STAGE,
-								&objects,
+								&objects, NULL,
 								concurrentCons,
 								progname, echo, quiet);
 	}
@@ -461,8 +467,36 @@ escape_quotes(const char *src)
 /*
  * vacuum_one_database
  *
- * Process tables in the given database.  If the 'objects' list is empty,
- * process all tables in the database.
+ * Process tables in the given database.
+ *
+ * There are two ways to specify the list of objects to process:
+ *
+ * 1) The "found_objs" parameter is a double pointer to a fully qualified list
+ *    of objects to process, as returned by a previous call to
+ *    vacuum_one_database().
+ *
+ *     a) If both "found_objs" (the double pointer) and "*found_objs" (the
+ *        once-dereferenced double pointer) are not NULL, this list takes
+ *        priority, and anything specified in "objects" is ignored.
+ *
+ *     b) If "found_objs" (the double pointer) is not NULL but "*found_objs"
+ *        (the once-dereferenced double pointer) _is_ NULL, the "objects"
+ *        parameter takes priority, and the results of the catalog query
+ *        described in (2) are stored in "found_objs".
+ *
+ *     c) If "found_objs" (the double pointer) is NULL, the "objects"
+ *        parameter again takes priority, and the results of the catalog query
+ *        are not saved.
+ *
+ * 2) The "objects" parameter is a user-specified list of objects to process.
+ *    When (1b) or (1c) applies, this function performs a catalog query to
+ *    retrieve a fully qualified list of objects to process, as described
+ *    below.
+ *
+ *     a) If "objects" is not NULL, the catalog query gathers only the objects
+ *        listed in "objects".
+ *
+ *     b) If "objects" is NULL, all tables in the database are gathered.
  *
  * Note that this function is only concerned with running exactly one stage
  * when in analyze-in-stages mode; caller must iterate on us if necessary.
@@ -475,22 +509,18 @@ vacuum_one_database(ConnParams *cparams,
 					vacuumingOptions *vacopts,
 					int stage,
 					SimpleStringList *objects,
+					SimpleStringList **found_objs,
 					int concurrentCons,
 					const char *progname, bool echo, bool quiet)
 {
 	PQExpBufferData sql;
-	PQExpBufferData buf;
-	PQExpBufferData catalog_query;
-	PGresult   *res;
 	PGconn	   *conn;
 	SimpleStringListCell *cell;
 	ParallelSlotArray *sa;
-	SimpleStringList dbtables = {NULL, NULL};
-	int			i;
-	int			ntups;
+	int			ntups = 0;
 	bool		failed = false;
-	bool		objects_listed = false;
 	const char *initcmd;
+	SimpleStringList *ret = NULL;
 	const char *stage_commands[] = {
 		"SET default_statistics_target=1; SET vacuum_cost_delay=0;",
 		"SET default_statistics_target=10; RESET vacuum_cost_delay;",
@@ -599,19 +629,155 @@ vacuum_one_database(ConnParams *cparams,
 	}
 
 	/*
-	 * Prepare the list of tables to process by querying the catalogs.
-	 *
-	 * Since we execute the constructed query with the default search_path
-	 * (which could be unsafe), everything in this query MUST be fully
-	 * qualified.
-	 *
-	 * First, build a WITH clause for the catalog query if any tables were
-	 * specified, with a set of values made of relation names and their
-	 * optional set of columns.  This is used to match any provided column
-	 * lists with the generated qualified identifiers and to filter for the
-	 * tables provided via --table.  If a listed table does not exist, the
-	 * catalog query will fail.
+	 * If the caller provided the results of a previous catalog query, just
+	 * use that.  Otherwise, run the catalog query ourselves and set the
+	 * return variable if provided.
+	 */
+	if (found_objs && *found_objs)
+		ret = *found_objs;
+	else
+	{
+		ret = retrieve_objects(conn, vacopts, objects, echo);
+		if (found_objs)
+			*found_objs = ret;
+	}
+
+	/*
+	 * Count the number of objects in the catalog query result.  If there are
+	 * none, we are done.
+	 */
+	for (cell = ret ? ret->head : NULL; cell; cell = cell->next)
+		ntups++;
+
+	if (ntups == 0)
+	{
+		PQfinish(conn);
+		return;
+	}
+
+	/*
+	 * Ensure concurrentCons is sane.  If there are more connections than
+	 * vacuumable relations, we don't need to use them all.
 	 */
+	if (concurrentCons > ntups)
+		concurrentCons = ntups;
+	if (concurrentCons <= 0)
+		concurrentCons = 1;
+
+	/*
+	 * All slots need to be prepared to run the appropriate analyze stage, if
+	 * caller requested that mode.  We have to prepare the initial connection
+	 * ourselves before setting up the slots.
+	 */
+	if (stage == ANALYZE_NO_STAGE)
+		initcmd = NULL;
+	else
+	{
+		initcmd = stage_commands[stage];
+		executeCommand(conn, initcmd, echo);
+	}
+
+	/*
+	 * Setup the database connections. We reuse the connection we already have
+	 * for the first slot.  If not in parallel mode, the first slot in the
+	 * array contains the connection.
+	 */
+	sa = ParallelSlotsSetup(concurrentCons, cparams, progname, echo, initcmd);
+	ParallelSlotsAdoptConn(sa, conn);
+
+	initPQExpBuffer(&sql);
+
+	cell = ret->head;
+	do
+	{
+		const char *tabname = cell->val;
+		ParallelSlot *free_slot;
+
+		if (CancelRequested)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		free_slot = ParallelSlotsGetIdle(sa, NULL);
+		if (!free_slot)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		prepare_vacuum_command(&sql, PQserverVersion(free_slot->connection),
+							   vacopts, tabname);
+
+		/*
+		 * Execute the vacuum.  All errors are handled in processQueryResult
+		 * through ParallelSlotsGetIdle.
+		 */
+		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
+		run_vacuum_command(free_slot->connection, sql.data,
+						   echo, tabname);
+
+		cell = cell->next;
+	} while (cell != NULL);
+
+	if (!ParallelSlotsWaitCompletion(sa))
+	{
+		failed = true;
+		goto finish;
+	}
+
+	/* If we used SKIP_DATABASE_STATS, mop up with ONLY_DATABASE_STATS */
+	if (vacopts->skip_database_stats && stage == ANALYZE_NO_STAGE)
+	{
+		const char *cmd = "VACUUM (ONLY_DATABASE_STATS);";
+		ParallelSlot *free_slot = ParallelSlotsGetIdle(sa, NULL);
+
+		if (!free_slot)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
+		run_vacuum_command(free_slot->connection, cmd, echo, NULL);
+
+		if (!ParallelSlotsWaitCompletion(sa))
+			failed = true;
+	}
+
+finish:
+	ParallelSlotsTerminate(sa);
+	pg_free(sa);
+
+	termPQExpBuffer(&sql);
+
+	if (failed)
+		exit(1);
+}
+
+/*
+ * Prepare the list of tables to process by querying the catalogs.
+ *
+ * Since we execute the constructed query with the default search_path (which
+ * could be unsafe), everything in this query MUST be fully qualified.
+ *
+ * First, build a WITH clause for the catalog query if any tables were
+ * specified, with a set of values made of relation names and their optional
+ * set of columns.  This is used to match any provided column lists with the
+ * generated qualified identifiers and to filter for the tables provided via
+ * --table.  If a listed table does not exist, the catalog query will fail.
+ */
+static SimpleStringList *
+retrieve_objects(PGconn *conn, vacuumingOptions *vacopts,
+				 SimpleStringList *objects, bool echo)
+{
+	PQExpBufferData buf;
+	PQExpBufferData catalog_query;
+	PGresult   *res;
+	SimpleStringListCell *cell;
+	SimpleStringList *found_objs = palloc0(sizeof(SimpleStringList));
+	bool		objects_listed = false;
+
 	initPQExpBuffer(&catalog_query);
 	for (cell = objects ? objects->head : NULL; cell; cell = cell->next)
 	{
@@ -765,23 +931,12 @@ vacuum_one_database(ConnParams *cparams,
 	termPQExpBuffer(&catalog_query);
 	PQclear(executeQuery(conn, ALWAYS_SECURE_SEARCH_PATH_SQL, echo));
 
-	/*
-	 * If no rows are returned, there are no matching tables, so we are done.
-	 */
-	ntups = PQntuples(res);
-	if (ntups == 0)
-	{
-		PQclear(res);
-		PQfinish(conn);
-		return;
-	}
-
 	/*
 	 * Build qualified identifiers for each table, including the column list
 	 * if given.
 	 */
 	initPQExpBuffer(&buf);
-	for (i = 0; i < ntups; i++)
+	for (int i = 0; i < PQntuples(res); i++)
 	{
 		appendPQExpBufferStr(&buf,
 							 fmtQualifiedIdEnc(PQgetvalue(res, i, 1),
@@ -791,110 +946,13 @@ vacuum_one_database(ConnParams *cparams,
 		if (objects_listed && !PQgetisnull(res, i, 2))
 			appendPQExpBufferStr(&buf, PQgetvalue(res, i, 2));
 
-		simple_string_list_append(&dbtables, buf.data);
+		simple_string_list_append(found_objs, buf.data);
 		resetPQExpBuffer(&buf);
 	}
 	termPQExpBuffer(&buf);
 	PQclear(res);
 
-	/*
-	 * Ensure concurrentCons is sane.  If there are more connections than
-	 * vacuumable relations, we don't need to use them all.
-	 */
-	if (concurrentCons > ntups)
-		concurrentCons = ntups;
-	if (concurrentCons <= 0)
-		concurrentCons = 1;
-
-	/*
-	 * All slots need to be prepared to run the appropriate analyze stage, if
-	 * caller requested that mode.  We have to prepare the initial connection
-	 * ourselves before setting up the slots.
-	 */
-	if (stage == ANALYZE_NO_STAGE)
-		initcmd = NULL;
-	else
-	{
-		initcmd = stage_commands[stage];
-		executeCommand(conn, initcmd, echo);
-	}
-
-	/*
-	 * Setup the database connections. We reuse the connection we already have
-	 * for the first slot.  If not in parallel mode, the first slot in the
-	 * array contains the connection.
-	 */
-	sa = ParallelSlotsSetup(concurrentCons, cparams, progname, echo, initcmd);
-	ParallelSlotsAdoptConn(sa, conn);
-
-	initPQExpBuffer(&sql);
-
-	cell = dbtables.head;
-	do
-	{
-		const char *tabname = cell->val;
-		ParallelSlot *free_slot;
-
-		if (CancelRequested)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		free_slot = ParallelSlotsGetIdle(sa, NULL);
-		if (!free_slot)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		prepare_vacuum_command(&sql, PQserverVersion(free_slot->connection),
-							   vacopts, tabname);
-
-		/*
-		 * Execute the vacuum.  All errors are handled in processQueryResult
-		 * through ParallelSlotsGetIdle.
-		 */
-		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
-		run_vacuum_command(free_slot->connection, sql.data,
-						   echo, tabname);
-
-		cell = cell->next;
-	} while (cell != NULL);
-
-	if (!ParallelSlotsWaitCompletion(sa))
-	{
-		failed = true;
-		goto finish;
-	}
-
-	/* If we used SKIP_DATABASE_STATS, mop up with ONLY_DATABASE_STATS */
-	if (vacopts->skip_database_stats && stage == ANALYZE_NO_STAGE)
-	{
-		const char *cmd = "VACUUM (ONLY_DATABASE_STATS);";
-		ParallelSlot *free_slot = ParallelSlotsGetIdle(sa, NULL);
-
-		if (!free_slot)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
-		run_vacuum_command(free_slot->connection, cmd, echo, NULL);
-
-		if (!ParallelSlotsWaitCompletion(sa))
-			failed = true;
-	}
-
-finish:
-	ParallelSlotsTerminate(sa);
-	pg_free(sa);
-
-	termPQExpBuffer(&sql);
-
-	if (failed)
-		exit(1);
+	return found_objs;
 }
 
 /*
@@ -941,7 +999,7 @@ vacuum_all_databases(ConnParams *cparams,
 
 				vacuum_one_database(cparams, vacopts,
 									stage,
-									objects,
+									objects, NULL,
 									concurrentCons,
 									progname, echo, quiet);
 			}
@@ -955,7 +1013,7 @@ vacuum_all_databases(ConnParams *cparams,
 
 			vacuum_one_database(cparams, vacopts,
 								ANALYZE_NO_STAGE,
-								objects,
+								objects, NULL,
 								concurrentCons,
 								progname, echo, quiet);
 		}
-- 
2.39.5 (Apple Git-154)


--B27oui4/QxrBRsZE
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment;
	filename="v6-0002-vacuumdb-Add-option-for-analyzing-only-relations-.patch"



^ permalink  raw  reply  [nested|flat] 4+ messages in thread

* [PATCH v5 1/3] vacuumdb: Teach vacuum_one_database() to reuse query results.
@ 2025-03-11 16:18 Nathan Bossart <[email protected]>
  0 siblings, 0 replies; 4+ messages in thread

From: Nathan Bossart @ 2025-03-11 16:18 UTC (permalink / raw)

Presently, each call to vacuum_one_database() performs a catalog
query to retrieve the list of tables to process.  A proposed
follow-up commit would add a "missing only" feature to
--analyze-in-stages, which requires us to save the results of the
catalog query (since tables without statistics would have them
after the first stage).  This commit adds this ability via a new
parameter for vacuum_one_database() that specifies either a
previously-retrieved list to process or a place to store the
results of the catalog query for later use.  Note that this commit
does not make use of this new parameter for anything.  That is left
as a future exercise.

The trade-offs of this approach are increased memory usage and less
responsiveness to concurrent catalog changes in later stages,
neither of which is expected to bother anyone.

Author: Corey Huinker <[email protected]>
Co-authored-by: Nathan Bossart <[email protected]>
Reviewed-by: John Naylor <[email protected]>
Discussion: https://postgr.es/m/Z5O1bpcwDrMgyrYy%40nathan
---
 src/bin/scripts/vacuumdb.c | 330 ++++++++++++++++++++++---------------
 1 file changed, 194 insertions(+), 136 deletions(-)

diff --git a/src/bin/scripts/vacuumdb.c b/src/bin/scripts/vacuumdb.c
index 982bf070be6..e28f82a0eba 100644
--- a/src/bin/scripts/vacuumdb.c
+++ b/src/bin/scripts/vacuumdb.c
@@ -62,10 +62,16 @@ typedef enum
 
 static VacObjFilter objfilter = OBJFILTER_NONE;
 
+static SimpleStringList *retrieve_objects(PGconn *conn,
+										  vacuumingOptions *vacopts,
+										  SimpleStringList *objects,
+										  bool echo);
+
 static void vacuum_one_database(ConnParams *cparams,
 								vacuumingOptions *vacopts,
 								int stage,
 								SimpleStringList *objects,
+								SimpleStringList **found_objs,
 								int concurrentCons,
 								const char *progname, bool echo, bool quiet);
 
@@ -405,7 +411,7 @@ main(int argc, char *argv[])
 			{
 				vacuum_one_database(&cparams, &vacopts,
 									stage,
-									&objects,
+									&objects, NULL,
 									concurrentCons,
 									progname, echo, quiet);
 			}
@@ -413,7 +419,7 @@ main(int argc, char *argv[])
 		else
 			vacuum_one_database(&cparams, &vacopts,
 								ANALYZE_NO_STAGE,
-								&objects,
+								&objects, NULL,
 								concurrentCons,
 								progname, echo, quiet);
 	}
@@ -461,8 +467,36 @@ escape_quotes(const char *src)
 /*
  * vacuum_one_database
  *
- * Process tables in the given database.  If the 'objects' list is empty,
- * process all tables in the database.
+ * Process tables in the given database.
+ *
+ * There are two ways to specify the list of objects to process:
+ *
+ * 1) The "found_objs" parameter is a double pointer to a fully qualified list
+ *    of objects to process, as returned by a previous call to
+ *    vacuum_one_database().
+ *
+ *     a) If both "found_objs" (the double pointer) and "*found_objs" (the
+ *        once-dereferenced double pointer) are not NULL, this list takes
+ *        priority, and anything specified in "objects" is ignored.
+ *
+ *     b) If "found_objs" (the double pointer) is not NULL but "*found_objs"
+ *        (the once-dereferenced double pointer) _is_ NULL, the "objects"
+ *        parameter takes priority, and the results of the catalog query
+ *        described in (2) are stored in "found_objs".
+ *
+ *     c) If "found_objs" (the double pointer) is NULL, the "objects"
+ *        parameter again takes priority, and the results of the catalog query
+ *        are not saved.
+ *
+ * 2) The "objects" parameter is a user-specified list of objects to process.
+ *    When (1b) or (1c) applies, this function performs a catalog query to
+ *    retrieve a fully qualified list of objects to process, as described
+ *    below.
+ *
+ *     a) If "objects" is not NULL, the catalog query gathers only the objects
+ *        listed in "objects".
+ *
+ *     b) If "objects" is NULL, all tables in the database are gathered.
  *
  * Note that this function is only concerned with running exactly one stage
  * when in analyze-in-stages mode; caller must iterate on us if necessary.
@@ -475,22 +509,18 @@ vacuum_one_database(ConnParams *cparams,
 					vacuumingOptions *vacopts,
 					int stage,
 					SimpleStringList *objects,
+					SimpleStringList **found_objs,
 					int concurrentCons,
 					const char *progname, bool echo, bool quiet)
 {
 	PQExpBufferData sql;
-	PQExpBufferData buf;
-	PQExpBufferData catalog_query;
-	PGresult   *res;
 	PGconn	   *conn;
 	SimpleStringListCell *cell;
 	ParallelSlotArray *sa;
-	SimpleStringList dbtables = {NULL, NULL};
-	int			i;
-	int			ntups;
+	int			ntups = 0;
 	bool		failed = false;
-	bool		objects_listed = false;
 	const char *initcmd;
+	SimpleStringList *ret = NULL;
 	const char *stage_commands[] = {
 		"SET default_statistics_target=1; SET vacuum_cost_delay=0;",
 		"SET default_statistics_target=10; RESET vacuum_cost_delay;",
@@ -599,19 +629,155 @@ vacuum_one_database(ConnParams *cparams,
 	}
 
 	/*
-	 * Prepare the list of tables to process by querying the catalogs.
-	 *
-	 * Since we execute the constructed query with the default search_path
-	 * (which could be unsafe), everything in this query MUST be fully
-	 * qualified.
-	 *
-	 * First, build a WITH clause for the catalog query if any tables were
-	 * specified, with a set of values made of relation names and their
-	 * optional set of columns.  This is used to match any provided column
-	 * lists with the generated qualified identifiers and to filter for the
-	 * tables provided via --table.  If a listed table does not exist, the
-	 * catalog query will fail.
+	 * If the caller provided the results of a previous catalog query, just
+	 * use that.  Otherwise, run the catalog query ourselves and set the
+	 * return variable if provided.
+	 */
+	if (found_objs && *found_objs)
+		ret = *found_objs;
+	else
+	{
+		ret = retrieve_objects(conn, vacopts, objects, echo);
+		if (found_objs)
+			*found_objs = ret;
+	}
+
+	/*
+	 * Count the number of objects in the catalog query result.  If there are
+	 * none, we are done.
+	 */
+	for (cell = ret ? ret->head : NULL; cell; cell = cell->next)
+		ntups++;
+
+	if (ntups == 0)
+	{
+		PQfinish(conn);
+		return;
+	}
+
+	/*
+	 * Ensure concurrentCons is sane.  If there are more connections than
+	 * vacuumable relations, we don't need to use them all.
 	 */
+	if (concurrentCons > ntups)
+		concurrentCons = ntups;
+	if (concurrentCons <= 0)
+		concurrentCons = 1;
+
+	/*
+	 * All slots need to be prepared to run the appropriate analyze stage, if
+	 * caller requested that mode.  We have to prepare the initial connection
+	 * ourselves before setting up the slots.
+	 */
+	if (stage == ANALYZE_NO_STAGE)
+		initcmd = NULL;
+	else
+	{
+		initcmd = stage_commands[stage];
+		executeCommand(conn, initcmd, echo);
+	}
+
+	/*
+	 * Setup the database connections. We reuse the connection we already have
+	 * for the first slot.  If not in parallel mode, the first slot in the
+	 * array contains the connection.
+	 */
+	sa = ParallelSlotsSetup(concurrentCons, cparams, progname, echo, initcmd);
+	ParallelSlotsAdoptConn(sa, conn);
+
+	initPQExpBuffer(&sql);
+
+	cell = ret->head;
+	do
+	{
+		const char *tabname = cell->val;
+		ParallelSlot *free_slot;
+
+		if (CancelRequested)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		free_slot = ParallelSlotsGetIdle(sa, NULL);
+		if (!free_slot)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		prepare_vacuum_command(&sql, PQserverVersion(free_slot->connection),
+							   vacopts, tabname);
+
+		/*
+		 * Execute the vacuum.  All errors are handled in processQueryResult
+		 * through ParallelSlotsGetIdle.
+		 */
+		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
+		run_vacuum_command(free_slot->connection, sql.data,
+						   echo, tabname);
+
+		cell = cell->next;
+	} while (cell != NULL);
+
+	if (!ParallelSlotsWaitCompletion(sa))
+	{
+		failed = true;
+		goto finish;
+	}
+
+	/* If we used SKIP_DATABASE_STATS, mop up with ONLY_DATABASE_STATS */
+	if (vacopts->skip_database_stats && stage == ANALYZE_NO_STAGE)
+	{
+		const char *cmd = "VACUUM (ONLY_DATABASE_STATS);";
+		ParallelSlot *free_slot = ParallelSlotsGetIdle(sa, NULL);
+
+		if (!free_slot)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
+		run_vacuum_command(free_slot->connection, cmd, echo, NULL);
+
+		if (!ParallelSlotsWaitCompletion(sa))
+			failed = true;
+	}
+
+finish:
+	ParallelSlotsTerminate(sa);
+	pg_free(sa);
+
+	termPQExpBuffer(&sql);
+
+	if (failed)
+		exit(1);
+}
+
+/*
+ * Prepare the list of tables to process by querying the catalogs.
+ *
+ * Since we execute the constructed query with the default search_path (which
+ * could be unsafe), everything in this query MUST be fully qualified.
+ *
+ * First, build a WITH clause for the catalog query if any tables were
+ * specified, with a set of values made of relation names and their optional
+ * set of columns.  This is used to match any provided column lists with the
+ * generated qualified identifiers and to filter for the tables provided via
+ * --table.  If a listed table does not exist, the catalog query will fail.
+ */
+static SimpleStringList *
+retrieve_objects(PGconn *conn, vacuumingOptions *vacopts,
+				 SimpleStringList *objects, bool echo)
+{
+	PQExpBufferData buf;
+	PQExpBufferData catalog_query;
+	PGresult   *res;
+	SimpleStringListCell *cell;
+	SimpleStringList *found_objs = palloc0(sizeof(SimpleStringList));
+	bool		objects_listed = false;
+
 	initPQExpBuffer(&catalog_query);
 	for (cell = objects ? objects->head : NULL; cell; cell = cell->next)
 	{
@@ -765,23 +931,12 @@ vacuum_one_database(ConnParams *cparams,
 	termPQExpBuffer(&catalog_query);
 	PQclear(executeQuery(conn, ALWAYS_SECURE_SEARCH_PATH_SQL, echo));
 
-	/*
-	 * If no rows are returned, there are no matching tables, so we are done.
-	 */
-	ntups = PQntuples(res);
-	if (ntups == 0)
-	{
-		PQclear(res);
-		PQfinish(conn);
-		return;
-	}
-
 	/*
 	 * Build qualified identifiers for each table, including the column list
 	 * if given.
 	 */
 	initPQExpBuffer(&buf);
-	for (i = 0; i < ntups; i++)
+	for (int i = 0; i < PQntuples(res); i++)
 	{
 		appendPQExpBufferStr(&buf,
 							 fmtQualifiedIdEnc(PQgetvalue(res, i, 1),
@@ -791,110 +946,13 @@ vacuum_one_database(ConnParams *cparams,
 		if (objects_listed && !PQgetisnull(res, i, 2))
 			appendPQExpBufferStr(&buf, PQgetvalue(res, i, 2));
 
-		simple_string_list_append(&dbtables, buf.data);
+		simple_string_list_append(found_objs, buf.data);
 		resetPQExpBuffer(&buf);
 	}
 	termPQExpBuffer(&buf);
 	PQclear(res);
 
-	/*
-	 * Ensure concurrentCons is sane.  If there are more connections than
-	 * vacuumable relations, we don't need to use them all.
-	 */
-	if (concurrentCons > ntups)
-		concurrentCons = ntups;
-	if (concurrentCons <= 0)
-		concurrentCons = 1;
-
-	/*
-	 * All slots need to be prepared to run the appropriate analyze stage, if
-	 * caller requested that mode.  We have to prepare the initial connection
-	 * ourselves before setting up the slots.
-	 */
-	if (stage == ANALYZE_NO_STAGE)
-		initcmd = NULL;
-	else
-	{
-		initcmd = stage_commands[stage];
-		executeCommand(conn, initcmd, echo);
-	}
-
-	/*
-	 * Setup the database connections. We reuse the connection we already have
-	 * for the first slot.  If not in parallel mode, the first slot in the
-	 * array contains the connection.
-	 */
-	sa = ParallelSlotsSetup(concurrentCons, cparams, progname, echo, initcmd);
-	ParallelSlotsAdoptConn(sa, conn);
-
-	initPQExpBuffer(&sql);
-
-	cell = dbtables.head;
-	do
-	{
-		const char *tabname = cell->val;
-		ParallelSlot *free_slot;
-
-		if (CancelRequested)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		free_slot = ParallelSlotsGetIdle(sa, NULL);
-		if (!free_slot)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		prepare_vacuum_command(&sql, PQserverVersion(free_slot->connection),
-							   vacopts, tabname);
-
-		/*
-		 * Execute the vacuum.  All errors are handled in processQueryResult
-		 * through ParallelSlotsGetIdle.
-		 */
-		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
-		run_vacuum_command(free_slot->connection, sql.data,
-						   echo, tabname);
-
-		cell = cell->next;
-	} while (cell != NULL);
-
-	if (!ParallelSlotsWaitCompletion(sa))
-	{
-		failed = true;
-		goto finish;
-	}
-
-	/* If we used SKIP_DATABASE_STATS, mop up with ONLY_DATABASE_STATS */
-	if (vacopts->skip_database_stats && stage == ANALYZE_NO_STAGE)
-	{
-		const char *cmd = "VACUUM (ONLY_DATABASE_STATS);";
-		ParallelSlot *free_slot = ParallelSlotsGetIdle(sa, NULL);
-
-		if (!free_slot)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
-		run_vacuum_command(free_slot->connection, cmd, echo, NULL);
-
-		if (!ParallelSlotsWaitCompletion(sa))
-			failed = true;
-	}
-
-finish:
-	ParallelSlotsTerminate(sa);
-	pg_free(sa);
-
-	termPQExpBuffer(&sql);
-
-	if (failed)
-		exit(1);
+	return found_objs;
 }
 
 /*
@@ -941,7 +999,7 @@ vacuum_all_databases(ConnParams *cparams,
 
 				vacuum_one_database(cparams, vacopts,
 									stage,
-									objects,
+									objects, NULL,
 									concurrentCons,
 									progname, echo, quiet);
 			}
@@ -955,7 +1013,7 @@ vacuum_all_databases(ConnParams *cparams,
 
 			vacuum_one_database(cparams, vacopts,
 								ANALYZE_NO_STAGE,
-								objects,
+								objects, NULL,
 								concurrentCons,
 								progname, echo, quiet);
 		}
-- 
2.39.5 (Apple Git-154)


--zUTpaz3RFkyl1aol
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment;
	filename="v5-0002-vacuumdb-Add-option-for-analyzing-only-relations-.patch"



^ permalink  raw  reply  [nested|flat] 4+ messages in thread

* [PATCH v7 1/3] vacuumdb: Teach vacuum_one_database() to reuse query results.
@ 2025-03-17 21:11 Nathan Bossart <[email protected]>
  0 siblings, 0 replies; 4+ messages in thread

From: Nathan Bossart @ 2025-03-17 21:11 UTC (permalink / raw)

Presently, each call to vacuum_one_database() performs a catalog
query to retrieve the list of tables to process.  A proposed
follow-up commit would add a "missing only" feature to
--analyze-in-stages, which requires us to save the results of the
catalog query (since tables without statistics would have them
after the first stage).  This commit adds this ability via a new
parameter for vacuum_one_database() that specifies either a
previously-retrieved list to process or a place to return the
results of the catalog query.  Note that this commit does not make
use of this new parameter for anything.  That is left for a
follow-up commit.

Author: Corey Huinker <[email protected]>
Co-authored-by: Nathan Bossart <[email protected]>
Reviewed-by: John Naylor <[email protected]>
Discussion: https://postgr.es/m/Z5O1bpcwDrMgyrYy%40nathan
---
 src/bin/scripts/vacuumdb.c | 330 ++++++++++++++++++++++---------------
 1 file changed, 194 insertions(+), 136 deletions(-)

diff --git a/src/bin/scripts/vacuumdb.c b/src/bin/scripts/vacuumdb.c
index 982bf070be6..e28f82a0eba 100644
--- a/src/bin/scripts/vacuumdb.c
+++ b/src/bin/scripts/vacuumdb.c
@@ -62,10 +62,16 @@ typedef enum
 
 static VacObjFilter objfilter = OBJFILTER_NONE;
 
+static SimpleStringList *retrieve_objects(PGconn *conn,
+										  vacuumingOptions *vacopts,
+										  SimpleStringList *objects,
+										  bool echo);
+
 static void vacuum_one_database(ConnParams *cparams,
 								vacuumingOptions *vacopts,
 								int stage,
 								SimpleStringList *objects,
+								SimpleStringList **found_objs,
 								int concurrentCons,
 								const char *progname, bool echo, bool quiet);
 
@@ -405,7 +411,7 @@ main(int argc, char *argv[])
 			{
 				vacuum_one_database(&cparams, &vacopts,
 									stage,
-									&objects,
+									&objects, NULL,
 									concurrentCons,
 									progname, echo, quiet);
 			}
@@ -413,7 +419,7 @@ main(int argc, char *argv[])
 		else
 			vacuum_one_database(&cparams, &vacopts,
 								ANALYZE_NO_STAGE,
-								&objects,
+								&objects, NULL,
 								concurrentCons,
 								progname, echo, quiet);
 	}
@@ -461,8 +467,36 @@ escape_quotes(const char *src)
 /*
  * vacuum_one_database
  *
- * Process tables in the given database.  If the 'objects' list is empty,
- * process all tables in the database.
+ * Process tables in the given database.
+ *
+ * There are two ways to specify the list of objects to process:
+ *
+ * 1) The "found_objs" parameter is a double pointer to a fully qualified list
+ *    of objects to process, as returned by a previous call to
+ *    vacuum_one_database().
+ *
+ *     a) If both "found_objs" (the double pointer) and "*found_objs" (the
+ *        once-dereferenced double pointer) are not NULL, this list takes
+ *        priority, and anything specified in "objects" is ignored.
+ *
+ *     b) If "found_objs" (the double pointer) is not NULL but "*found_objs"
+ *        (the once-dereferenced double pointer) _is_ NULL, the "objects"
+ *        parameter takes priority, and the results of the catalog query
+ *        described in (2) are stored in "found_objs".
+ *
+ *     c) If "found_objs" (the double pointer) is NULL, the "objects"
+ *        parameter again takes priority, and the results of the catalog query
+ *        are not saved.
+ *
+ * 2) The "objects" parameter is a user-specified list of objects to process.
+ *    When (1b) or (1c) applies, this function performs a catalog query to
+ *    retrieve a fully qualified list of objects to process, as described
+ *    below.
+ *
+ *     a) If "objects" is not NULL, the catalog query gathers only the objects
+ *        listed in "objects".
+ *
+ *     b) If "objects" is NULL, all tables in the database are gathered.
  *
  * Note that this function is only concerned with running exactly one stage
  * when in analyze-in-stages mode; caller must iterate on us if necessary.
@@ -475,22 +509,18 @@ vacuum_one_database(ConnParams *cparams,
 					vacuumingOptions *vacopts,
 					int stage,
 					SimpleStringList *objects,
+					SimpleStringList **found_objs,
 					int concurrentCons,
 					const char *progname, bool echo, bool quiet)
 {
 	PQExpBufferData sql;
-	PQExpBufferData buf;
-	PQExpBufferData catalog_query;
-	PGresult   *res;
 	PGconn	   *conn;
 	SimpleStringListCell *cell;
 	ParallelSlotArray *sa;
-	SimpleStringList dbtables = {NULL, NULL};
-	int			i;
-	int			ntups;
+	int			ntups = 0;
 	bool		failed = false;
-	bool		objects_listed = false;
 	const char *initcmd;
+	SimpleStringList *ret = NULL;
 	const char *stage_commands[] = {
 		"SET default_statistics_target=1; SET vacuum_cost_delay=0;",
 		"SET default_statistics_target=10; RESET vacuum_cost_delay;",
@@ -599,19 +629,155 @@ vacuum_one_database(ConnParams *cparams,
 	}
 
 	/*
-	 * Prepare the list of tables to process by querying the catalogs.
-	 *
-	 * Since we execute the constructed query with the default search_path
-	 * (which could be unsafe), everything in this query MUST be fully
-	 * qualified.
-	 *
-	 * First, build a WITH clause for the catalog query if any tables were
-	 * specified, with a set of values made of relation names and their
-	 * optional set of columns.  This is used to match any provided column
-	 * lists with the generated qualified identifiers and to filter for the
-	 * tables provided via --table.  If a listed table does not exist, the
-	 * catalog query will fail.
+	 * If the caller provided the results of a previous catalog query, just
+	 * use that.  Otherwise, run the catalog query ourselves and set the
+	 * return variable if provided.
+	 */
+	if (found_objs && *found_objs)
+		ret = *found_objs;
+	else
+	{
+		ret = retrieve_objects(conn, vacopts, objects, echo);
+		if (found_objs)
+			*found_objs = ret;
+	}
+
+	/*
+	 * Count the number of objects in the catalog query result.  If there are
+	 * none, we are done.
+	 */
+	for (cell = ret ? ret->head : NULL; cell; cell = cell->next)
+		ntups++;
+
+	if (ntups == 0)
+	{
+		PQfinish(conn);
+		return;
+	}
+
+	/*
+	 * Ensure concurrentCons is sane.  If there are more connections than
+	 * vacuumable relations, we don't need to use them all.
 	 */
+	if (concurrentCons > ntups)
+		concurrentCons = ntups;
+	if (concurrentCons <= 0)
+		concurrentCons = 1;
+
+	/*
+	 * All slots need to be prepared to run the appropriate analyze stage, if
+	 * caller requested that mode.  We have to prepare the initial connection
+	 * ourselves before setting up the slots.
+	 */
+	if (stage == ANALYZE_NO_STAGE)
+		initcmd = NULL;
+	else
+	{
+		initcmd = stage_commands[stage];
+		executeCommand(conn, initcmd, echo);
+	}
+
+	/*
+	 * Setup the database connections. We reuse the connection we already have
+	 * for the first slot.  If not in parallel mode, the first slot in the
+	 * array contains the connection.
+	 */
+	sa = ParallelSlotsSetup(concurrentCons, cparams, progname, echo, initcmd);
+	ParallelSlotsAdoptConn(sa, conn);
+
+	initPQExpBuffer(&sql);
+
+	cell = ret->head;
+	do
+	{
+		const char *tabname = cell->val;
+		ParallelSlot *free_slot;
+
+		if (CancelRequested)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		free_slot = ParallelSlotsGetIdle(sa, NULL);
+		if (!free_slot)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		prepare_vacuum_command(&sql, PQserverVersion(free_slot->connection),
+							   vacopts, tabname);
+
+		/*
+		 * Execute the vacuum.  All errors are handled in processQueryResult
+		 * through ParallelSlotsGetIdle.
+		 */
+		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
+		run_vacuum_command(free_slot->connection, sql.data,
+						   echo, tabname);
+
+		cell = cell->next;
+	} while (cell != NULL);
+
+	if (!ParallelSlotsWaitCompletion(sa))
+	{
+		failed = true;
+		goto finish;
+	}
+
+	/* If we used SKIP_DATABASE_STATS, mop up with ONLY_DATABASE_STATS */
+	if (vacopts->skip_database_stats && stage == ANALYZE_NO_STAGE)
+	{
+		const char *cmd = "VACUUM (ONLY_DATABASE_STATS);";
+		ParallelSlot *free_slot = ParallelSlotsGetIdle(sa, NULL);
+
+		if (!free_slot)
+		{
+			failed = true;
+			goto finish;
+		}
+
+		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
+		run_vacuum_command(free_slot->connection, cmd, echo, NULL);
+
+		if (!ParallelSlotsWaitCompletion(sa))
+			failed = true;
+	}
+
+finish:
+	ParallelSlotsTerminate(sa);
+	pg_free(sa);
+
+	termPQExpBuffer(&sql);
+
+	if (failed)
+		exit(1);
+}
+
+/*
+ * Prepare the list of tables to process by querying the catalogs.
+ *
+ * Since we execute the constructed query with the default search_path (which
+ * could be unsafe), everything in this query MUST be fully qualified.
+ *
+ * First, build a WITH clause for the catalog query if any tables were
+ * specified, with a set of values made of relation names and their optional
+ * set of columns.  This is used to match any provided column lists with the
+ * generated qualified identifiers and to filter for the tables provided via
+ * --table.  If a listed table does not exist, the catalog query will fail.
+ */
+static SimpleStringList *
+retrieve_objects(PGconn *conn, vacuumingOptions *vacopts,
+				 SimpleStringList *objects, bool echo)
+{
+	PQExpBufferData buf;
+	PQExpBufferData catalog_query;
+	PGresult   *res;
+	SimpleStringListCell *cell;
+	SimpleStringList *found_objs = palloc0(sizeof(SimpleStringList));
+	bool		objects_listed = false;
+
 	initPQExpBuffer(&catalog_query);
 	for (cell = objects ? objects->head : NULL; cell; cell = cell->next)
 	{
@@ -765,23 +931,12 @@ vacuum_one_database(ConnParams *cparams,
 	termPQExpBuffer(&catalog_query);
 	PQclear(executeQuery(conn, ALWAYS_SECURE_SEARCH_PATH_SQL, echo));
 
-	/*
-	 * If no rows are returned, there are no matching tables, so we are done.
-	 */
-	ntups = PQntuples(res);
-	if (ntups == 0)
-	{
-		PQclear(res);
-		PQfinish(conn);
-		return;
-	}
-
 	/*
 	 * Build qualified identifiers for each table, including the column list
 	 * if given.
 	 */
 	initPQExpBuffer(&buf);
-	for (i = 0; i < ntups; i++)
+	for (int i = 0; i < PQntuples(res); i++)
 	{
 		appendPQExpBufferStr(&buf,
 							 fmtQualifiedIdEnc(PQgetvalue(res, i, 1),
@@ -791,110 +946,13 @@ vacuum_one_database(ConnParams *cparams,
 		if (objects_listed && !PQgetisnull(res, i, 2))
 			appendPQExpBufferStr(&buf, PQgetvalue(res, i, 2));
 
-		simple_string_list_append(&dbtables, buf.data);
+		simple_string_list_append(found_objs, buf.data);
 		resetPQExpBuffer(&buf);
 	}
 	termPQExpBuffer(&buf);
 	PQclear(res);
 
-	/*
-	 * Ensure concurrentCons is sane.  If there are more connections than
-	 * vacuumable relations, we don't need to use them all.
-	 */
-	if (concurrentCons > ntups)
-		concurrentCons = ntups;
-	if (concurrentCons <= 0)
-		concurrentCons = 1;
-
-	/*
-	 * All slots need to be prepared to run the appropriate analyze stage, if
-	 * caller requested that mode.  We have to prepare the initial connection
-	 * ourselves before setting up the slots.
-	 */
-	if (stage == ANALYZE_NO_STAGE)
-		initcmd = NULL;
-	else
-	{
-		initcmd = stage_commands[stage];
-		executeCommand(conn, initcmd, echo);
-	}
-
-	/*
-	 * Setup the database connections. We reuse the connection we already have
-	 * for the first slot.  If not in parallel mode, the first slot in the
-	 * array contains the connection.
-	 */
-	sa = ParallelSlotsSetup(concurrentCons, cparams, progname, echo, initcmd);
-	ParallelSlotsAdoptConn(sa, conn);
-
-	initPQExpBuffer(&sql);
-
-	cell = dbtables.head;
-	do
-	{
-		const char *tabname = cell->val;
-		ParallelSlot *free_slot;
-
-		if (CancelRequested)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		free_slot = ParallelSlotsGetIdle(sa, NULL);
-		if (!free_slot)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		prepare_vacuum_command(&sql, PQserverVersion(free_slot->connection),
-							   vacopts, tabname);
-
-		/*
-		 * Execute the vacuum.  All errors are handled in processQueryResult
-		 * through ParallelSlotsGetIdle.
-		 */
-		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
-		run_vacuum_command(free_slot->connection, sql.data,
-						   echo, tabname);
-
-		cell = cell->next;
-	} while (cell != NULL);
-
-	if (!ParallelSlotsWaitCompletion(sa))
-	{
-		failed = true;
-		goto finish;
-	}
-
-	/* If we used SKIP_DATABASE_STATS, mop up with ONLY_DATABASE_STATS */
-	if (vacopts->skip_database_stats && stage == ANALYZE_NO_STAGE)
-	{
-		const char *cmd = "VACUUM (ONLY_DATABASE_STATS);";
-		ParallelSlot *free_slot = ParallelSlotsGetIdle(sa, NULL);
-
-		if (!free_slot)
-		{
-			failed = true;
-			goto finish;
-		}
-
-		ParallelSlotSetHandler(free_slot, TableCommandResultHandler, NULL);
-		run_vacuum_command(free_slot->connection, cmd, echo, NULL);
-
-		if (!ParallelSlotsWaitCompletion(sa))
-			failed = true;
-	}
-
-finish:
-	ParallelSlotsTerminate(sa);
-	pg_free(sa);
-
-	termPQExpBuffer(&sql);
-
-	if (failed)
-		exit(1);
+	return found_objs;
 }
 
 /*
@@ -941,7 +999,7 @@ vacuum_all_databases(ConnParams *cparams,
 
 				vacuum_one_database(cparams, vacopts,
 									stage,
-									objects,
+									objects, NULL,
 									concurrentCons,
 									progname, echo, quiet);
 			}
@@ -955,7 +1013,7 @@ vacuum_all_databases(ConnParams *cparams,
 
 			vacuum_one_database(cparams, vacopts,
 								ANALYZE_NO_STAGE,
-								objects,
+								objects, NULL,
 								concurrentCons,
 								progname, echo, quiet);
 		}
-- 
2.39.5 (Apple Git-154)


--GMqUTp9yliU3ROhG
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment;
	filename="v7-0002-vacuumdb-Add-option-for-analyzing-only-relations-.patch"



^ permalink  raw  reply  [nested|flat] 4+ messages in thread


end of thread, other threads:[~2025-03-17 21:11 UTC | newest]

Thread overview: 4+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2020-03-09 03:57 [PATCH v28 04/11] pg_ls_tmpdir to show directories and "isdir" argument.. Justin Pryzby <[email protected]>
2025-03-11 16:18 [PATCH v6 1/3] vacuumdb: Teach vacuum_one_database() to reuse query results. Nathan Bossart <[email protected]>
2025-03-11 16:18 [PATCH v5 1/3] vacuumdb: Teach vacuum_one_database() to reuse query results. Nathan Bossart <[email protected]>
2025-03-17 21:11 [PATCH v7 1/3] vacuumdb: Teach vacuum_one_database() to reuse query results. Nathan Bossart <[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