From d2ca9a50b5b99ef29aa65b68b5e6ddb253fbb04a Mon Sep 17 00:00:00 2001 From: David Rowley Date: Tue, 20 Feb 2024 18:34:21 +1300 Subject: [PATCH] Minor corrections for partition pruning When the partition pruning code finds an OpExpr with an operator that does not belong to the partition key's opfamily, the code checks to see if the negator of the operator is the opfamily's BTEqualStrategyNumber operator so that partition pruning can support that operator and invert the matching partitions. Doing this only works for LIST partitioned tables. Here we fix a minor correctness issue where when we discover we're not pruning for a LIST partitioned table, we return PARTCLAUSE_NOMATCH. PARTCLAUSE_NOMATCH is only meant to be used when the clause may match another partitioned key column. For this case, the clause is not going to be any more useful to another partitioned key as the partition strategy is not going to change from one key to the next. Noticed while working 4c2369ac5. No backpatch because returning PARTCLAUSE_NOMATCH instead of PARTCLAUSE_UNSUPPORTED mostly just causes wasted effort checking subsequent partition keys against a clause that will never be used for pruning. In passing, correct a comment for get_matching_range_bounds() which mentions that an 'opstrategy' of 0 is supported. It's not, so fix the comment. This was pointed out by Alexander Lakhin. Discussion: https://postgr.es/m/CAApHDvqriy8mPOFJ_Bd66YGXJ4+XULpv-4YdB+ePdCQFztyisA@mail.gmail.com Discussion: https://postgr.es/m/312fb507-9b5e-cf83-d8ed-cd0da72a902c@gmail.com --- src/backend/partitioning/partprune.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/src/backend/partitioning/partprune.c b/src/backend/partitioning/partprune.c index ccb9a579b30..a19f596d4de 100644 --- a/src/backend/partitioning/partprune.c +++ b/src/backend/partitioning/partprune.c @@ -1936,9 +1936,11 @@ match_clause_to_partition_key(GeneratePruningStepsContext *context, * whatsoever, but their negators (equality) are. We can use one of * those if we find it, but only for list partitioning. * - * Note: we report NOMATCH on failure, in case a later partkey has the - * same expression but different opfamily. That's unlikely, but not - * much more so than duplicate expressions with different collations. + * Note: we report NOMATCH on failure if the negator isn't the + * equality operator for the partkey's opfamily as other partkeys may + * have the same expression but different opfamily. That's unlikely, + * but not much more so than duplicate expressions with different + * collations. */ if (op_in_opfamily(opno, partopfamily)) { @@ -1948,8 +1950,9 @@ match_clause_to_partition_key(GeneratePruningStepsContext *context, } else { + /* not supported for anything apart from LIST partitioned tables */ if (part_scheme->strategy != PARTITION_STRATEGY_LIST) - return PARTCLAUSE_NOMATCH; + return PARTCLAUSE_UNSUPPORTED; /* See if the negator is equality */ negator = get_negator(opno); @@ -2924,7 +2927,7 @@ get_matching_list_bounds(PartitionPruneContext *context, * multiple pruning steps might exclude it, so we infer its inclusion * elsewhere. * - * 'opstrategy' if non-zero must be a btree strategy number. + * 'opstrategy' must be a btree strategy number. * * 'values' contains Datums indexed by the partition key to use for pruning. *