From 7654eb8e59e8fe5e918e517aa64ebeb5767da87c Mon Sep 17 00:00:00 2001 From: Tomas Vondra Date: Mon, 17 Jun 2024 16:48:30 +0200 Subject: [PATCH v20240617 43/56] Fix error "unexpected system attribute" when join with system attr We can't just change 'elog(ERROR, "unexpected system attribute");' to 'continue' in extract_relation_info since after we extract the StatisticExtInfo, and stat is not NULL, we grantee the expression in JoinPairInfo.clause has a matched expression with mcv_match_expression, however this is not true for system attribute. so fix it at the first place when populate the clause into JoinPairInfo.clauses which is the statext_is_supported_join_clause function. Expression contains a system attribute is OK since due to the implementation of mcv_match_expression so only Var need to be handled there. --- src/backend/statistics/extended_stats.c | 27 ++++++++++++++++++++----- src/test/regress/sql/stats_ext.sql | 6 ++++++ 2 files changed, 28 insertions(+), 5 deletions(-) diff --git a/src/backend/statistics/extended_stats.c b/src/backend/statistics/extended_stats.c index 08998f219e1..d6f1c70ae64 100644 --- a/src/backend/statistics/extended_stats.c +++ b/src/backend/statistics/extended_stats.c @@ -2846,11 +2846,11 @@ statext_determine_join_restrictions(PlannerInfo *root, RelOptInfo *rel, static bool statext_is_supported_join_clause(PlannerInfo *root, Node *clause) { - Oid oprsel; - RestrictInfo *rinfo; - OpExpr *opclause; - int left_relid, - right_relid; + Oid oprsel; + RestrictInfo *rinfo; + OpExpr *opclause; + int left_relid, right_relid; + Var *var; /* XXX Can we rely on always getting RestrictInfo here? */ if (!IsA(clause, RestrictInfo)) @@ -2908,6 +2908,14 @@ statext_is_supported_join_clause(PlannerInfo *root, Node *clause) if (left_relid == right_relid) return false; + var = (Var *) linitial(opclause->args); + if (IsA(var, Var) && var->varattno < 0) + return false; + + var = (Var *) lsecond(opclause->args); + if (IsA(var, Var) && var->varattno < 0) + return false; + return true; } @@ -3195,6 +3203,15 @@ extract_relation_info(PlannerInfo *root, JoinPairInfo *info, int index, } } + /* + * Find a stat which covers *all* the attnums and exprs for simplification. + * + * To overcome above limitation, statext_find_matching_mcv has to smart enough to + * decide which expression to discard as the first step. and later the other + * side of join has to use a stats which match or superset of expression here. + * at last mcv_combine_extended should be improved to handle the not-exactly-same + * mcv. + */ *stat = statext_find_matching_mcv(root, rel, attnums, exprs, base_conditions); return rel; diff --git a/src/test/regress/sql/stats_ext.sql b/src/test/regress/sql/stats_ext.sql index e372fffebfb..798ee78265e 100644 --- a/src/test/regress/sql/stats_ext.sql +++ b/src/test/regress/sql/stats_ext.sql @@ -1595,6 +1595,12 @@ SELECT * FROM check_estimated_rows('select * from join_test_1 j1 join join_test_ SELECT * FROM check_estimated_rows('select * from join_test_1 j1 join join_test_2 j2 on (j1.a + 1 = j2.a) where j1.c < 5'); SELECT * FROM check_estimated_rows('select * from join_test_1 j1 join join_test_2 j2 on (j1.a + 1 = j2.a + 1) where j1.c < 5'); +-- test join with system column var, but the ext statistics can't be built in system attribute AND extended statistics +-- must covers all the join columns, so the following 2 statements can use extended statistics for join. +SELECT * FROM check_estimated_rows('select * from join_test_1 j1 join join_test_2 j2 on ((j1.a + 1 = j2.a + 1) and (j1.b = j2.b)) and j1.cmin = j2.cmin'); +-- Join with system column expression. +SELECT * FROM check_estimated_rows('select * from join_test_1 j1 join join_test_2 j2 on ((j1.a + 1 = j2.a + 1) and (j1.b = j2.b)) and j1.cmin::text::int4 = j2.cmin::text::int4'); + -- try combining with single-column (and single-expression) statistics DROP STATISTICS join_stats_2; -- 2.45.2