From 94b6631986fad7d953f7ae9c14eabe077fe2be7d Mon Sep 17 00:00:00 2001 From: David Rowley Date: Sun, 23 May 2021 23:06:38 +1200 Subject: [PATCH v4 1/2] Track non-pruned partitions in RelOptInfo For partitioned tables with large numbers of partitions where queries are able to prune all but a very small number of partitions, as is common in partitioned tables used in OLTP workloads, the time spent in the planner looping over RelOptInfo.part_rels checking for non-NULL RelOptInfos could become a large portion of the overall planning time. Here we add a Bitmapset which records the non-pruned partitions. This allows us to more efficiently skip the pruned partitions by looping over the Bitmapset. This will very slightly slow down cases where no or not many partitions could be pruned, however, those cases are already slow to plan anyway and the overhead of looping over the Bitmapset would be unmeasurable when compared with the other duties the planner must perform. --- src/backend/nodes/outfuncs.c | 1 + src/backend/optimizer/path/joinrels.c | 1 + src/backend/optimizer/plan/planner.c | 31 ++++++++++++++++----------- src/backend/optimizer/util/inherit.c | 2 +- src/backend/optimizer/util/relnode.c | 3 +++ src/backend/partitioning/partprune.c | 7 +++--- src/include/nodes/pathnodes.h | 3 +++ 7 files changed, 31 insertions(+), 17 deletions(-) diff --git a/src/backend/nodes/outfuncs.c b/src/backend/nodes/outfuncs.c index 48202d2232..87561cbb6f 100644 --- a/src/backend/nodes/outfuncs.c +++ b/src/backend/nodes/outfuncs.c @@ -2386,6 +2386,7 @@ _outRelOptInfo(StringInfo str, const RelOptInfo *node) WRITE_BOOL_FIELD(consider_partitionwise_join); WRITE_BITMAPSET_FIELD(top_parent_relids); WRITE_BOOL_FIELD(partbounds_merged); + WRITE_BITMAPSET_FIELD(live_parts); WRITE_BITMAPSET_FIELD(all_partrels); } diff --git a/src/backend/optimizer/path/joinrels.c b/src/backend/optimizer/path/joinrels.c index 0dbe2ac726..8b69870cf4 100644 --- a/src/backend/optimizer/path/joinrels.c +++ b/src/backend/optimizer/path/joinrels.c @@ -1539,6 +1539,7 @@ try_partitionwise_join(PlannerInfo *root, RelOptInfo *rel1, RelOptInfo *rel2, child_sjinfo, child_sjinfo->jointype); joinrel->part_rels[cnt_parts] = child_joinrel; + joinrel->live_parts = bms_add_member(joinrel->live_parts, cnt_parts); joinrel->all_partrels = bms_add_members(joinrel->all_partrels, child_joinrel->relids); } diff --git a/src/backend/optimizer/plan/planner.c b/src/backend/optimizer/plan/planner.c index 86816ffe19..2cd691191c 100644 --- a/src/backend/optimizer/plan/planner.c +++ b/src/backend/optimizer/plan/planner.c @@ -6989,19 +6989,22 @@ apply_scanjoin_target_to_paths(PlannerInfo *root, if (rel_is_partitioned) { List *live_children = NIL; - int partition_idx; + int i; /* Adjust each partition. */ - for (partition_idx = 0; partition_idx < rel->nparts; partition_idx++) + i = -1; + while ((i = bms_next_member(rel->live_parts, i)) >= 0) { - RelOptInfo *child_rel = rel->part_rels[partition_idx]; + RelOptInfo *child_rel = rel->part_rels[i]; AppendRelInfo **appinfos; int nappinfos; List *child_scanjoin_targets = NIL; ListCell *lc; - /* Pruned or dummy children can be ignored. */ - if (child_rel == NULL || IS_DUMMY_REL(child_rel)) + Assert(child_rel != NULL); + + /* Dummy children can be ignored. */ + if (IS_DUMMY_REL(child_rel)) continue; /* Translate scan/join targets for this child. */ @@ -7082,32 +7085,36 @@ create_partitionwise_grouping_paths(PlannerInfo *root, PartitionwiseAggregateType patype, GroupPathExtraData *extra) { - int nparts = input_rel->nparts; - int cnt_parts; List *grouped_live_children = NIL; List *partially_grouped_live_children = NIL; PathTarget *target = grouped_rel->reltarget; bool partial_grouping_valid = true; + int i; Assert(patype != PARTITIONWISE_AGGREGATE_NONE); Assert(patype != PARTITIONWISE_AGGREGATE_PARTIAL || partially_grouped_rel != NULL); /* Add paths for partitionwise aggregation/grouping. */ - for (cnt_parts = 0; cnt_parts < nparts; cnt_parts++) + i = -1; + while ((i = bms_next_member(input_rel->live_parts, i)) >= 0) { - RelOptInfo *child_input_rel = input_rel->part_rels[cnt_parts]; - PathTarget *child_target = copy_pathtarget(target); + RelOptInfo *child_input_rel = input_rel->part_rels[i]; + PathTarget *child_target; AppendRelInfo **appinfos; int nappinfos; GroupPathExtraData child_extra; RelOptInfo *child_grouped_rel; RelOptInfo *child_partially_grouped_rel; - /* Pruned or dummy children can be ignored. */ - if (child_input_rel == NULL || IS_DUMMY_REL(child_input_rel)) + Assert(child_input_rel != NULL); + + /* Dummy children can be ignored. */ + if (IS_DUMMY_REL(child_input_rel)) continue; + child_target = copy_pathtarget(target); + /* * Copy the given "extra" structure as is and then override the * members specific to this child. diff --git a/src/backend/optimizer/util/inherit.c b/src/backend/optimizer/util/inherit.c index 992ef87b9d..c758172efa 100644 --- a/src/backend/optimizer/util/inherit.c +++ b/src/backend/optimizer/util/inherit.c @@ -348,7 +348,7 @@ expand_partitioned_rtentry(PlannerInfo *root, RelOptInfo *relinfo, * that survive pruning. Below, we will initialize child objects for the * surviving partitions. */ - live_parts = prune_append_rel_partitions(relinfo); + relinfo->live_parts = live_parts = prune_append_rel_partitions(relinfo); /* Expand simple_rel_array and friends to hold child objects. */ num_live_parts = bms_num_members(live_parts); diff --git a/src/backend/optimizer/util/relnode.c b/src/backend/optimizer/util/relnode.c index e105a4d5f1..47769cea45 100644 --- a/src/backend/optimizer/util/relnode.c +++ b/src/backend/optimizer/util/relnode.c @@ -255,6 +255,7 @@ build_simple_rel(PlannerInfo *root, int relid, RelOptInfo *parent) rel->partbounds_merged = false; rel->partition_qual = NIL; rel->part_rels = NULL; + rel->live_parts = NULL; rel->all_partrels = NULL; rel->partexprs = NULL; rel->nullable_partexprs = NULL; @@ -669,6 +670,7 @@ build_join_rel(PlannerInfo *root, joinrel->partbounds_merged = false; joinrel->partition_qual = NIL; joinrel->part_rels = NULL; + joinrel->live_parts = NULL; joinrel->all_partrels = NULL; joinrel->partexprs = NULL; joinrel->nullable_partexprs = NULL; @@ -847,6 +849,7 @@ build_child_join_rel(PlannerInfo *root, RelOptInfo *outer_rel, joinrel->partbounds_merged = false; joinrel->partition_qual = NIL; joinrel->part_rels = NULL; + joinrel->live_parts = NULL; joinrel->all_partrels = NULL; joinrel->partexprs = NULL; joinrel->nullable_partexprs = NULL; diff --git a/src/backend/partitioning/partprune.c b/src/backend/partitioning/partprune.c index c79374265c..e00edbe5c8 100644 --- a/src/backend/partitioning/partprune.c +++ b/src/backend/partitioning/partprune.c @@ -654,15 +654,14 @@ make_partitionedrel_pruneinfo(PlannerInfo *root, RelOptInfo *parentrel, relid_map = (Oid *) palloc0(nparts * sizeof(Oid)); present_parts = NULL; - for (i = 0; i < nparts; i++) + i = -1; + while ((i = bms_next_member(subpart->live_parts, i)) >= 0) { RelOptInfo *partrel = subpart->part_rels[i]; int subplanidx; int subpartidx; - /* Skip processing pruned partitions. */ - if (partrel == NULL) - continue; + Assert(partrel != NULL); subplan_map[i] = subplanidx = relid_subplan_map[partrel->relid] - 1; subpart_map[i] = subpartidx = relid_subpart_map[partrel->relid] - 1; diff --git a/src/include/nodes/pathnodes.h b/src/include/nodes/pathnodes.h index e20c245f98..ce7908c326 100644 --- a/src/include/nodes/pathnodes.h +++ b/src/include/nodes/pathnodes.h @@ -762,6 +762,9 @@ typedef struct RelOptInfo List *partition_qual; /* Partition constraint, if not the root */ struct RelOptInfo **part_rels; /* Array of RelOptInfos of partitions, * stored in the same order as bounds */ + Bitmapset *live_parts; /* Bitmap with members acting as indexes into + * the part_rels[] array to indicate which + * partitions survived partition pruning. */ Relids all_partrels; /* Relids set of all partition relids */ List **partexprs; /* Non-nullable partition key expressions */ List **nullable_partexprs; /* Nullable partition key expressions */ -- 2.30.2