Skip to content

Commit e5f5238

Browse files
danolivodanolivo
committed
Consider ndistinct on the first column in cost_sort().
Task: 9578. Tags: optimized_group_by.
1 parent 9c3dab9 commit e5f5238

File tree

3 files changed

+52
-25
lines changed

3 files changed

+52
-25
lines changed

‎src/backend/optimizer/path/costsize.c

Lines changed: 32 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -194,6 +194,8 @@ static double page_size(double tuples, int width);
194194
static double get_parallel_divisor(Path *path);
195195
static EquivalenceMember *identify_sort_ecmember(PlannerInfo *root,
196196
PathKey *key);
197+
static double sort_comparisons_factor(PlannerInfo *root, List *pathkeys,
198+
double ntuples);
197199

198200

199201
/*
@@ -2113,8 +2115,7 @@ cost_sort(Path *path, PlannerInfo *root,
21132115
{
21142116
Cost startup_cost;
21152117
Cost run_cost;
2116-
double cmpMultiplier =
2117-
(pathkeys == NIL) ? 2.0 : list_length(pathkeys) + 1.0;
2118+
double cmpMultiplier = sort_comparisons_factor(root, pathkeys, tuples);
21182119

21192120
cost_tuplesort(&startup_cost, &run_cost,
21202121
tuples, width, cmpMultiplier,
@@ -6560,4 +6561,32 @@ identify_sort_ecmember(PlannerInfo *root, PathKey *key)
65606561

65616562
Assert(candidate != NULL);
65626563
return candidate;
6563-
}
6564+
}
6565+
6566+
static double
6567+
sort_comparisons_factor(PlannerInfo *root, List *pathkeys, double ntuples)
6568+
{
6569+
int n = list_length(pathkeys);
6570+
double cmpMultiplier = (n == 0) ? 2.0 : n + 1;
6571+
6572+
if (root != NULL && ntuples > 1 && n > 1)
6573+
{
6574+
PathKey *key = linitial_node(PathKey, pathkeys);
6575+
EquivalenceMember *em = identify_sort_ecmember(root, key);
6576+
6577+
Assert(em->em_ndistinct >= 0);
6578+
6579+
if (em->em_ndistinct == 0.)
6580+
/*
6581+
* Optimiser doesn't have an info on ndistinct value, return
6582+
* extreme case
6583+
*/
6584+
return cmpMultiplier;
6585+
6586+
if (ntuples >= em->em_ndistinct)
6587+
cmpMultiplier =
6588+
2.0 + ((ntuples - em->em_ndistinct) / (ntuples - 1)) * (n - 1);
6589+
}
6590+
6591+
return cmpMultiplier;
6592+
}

‎src/test/regress/expected/aggregates.out

Lines changed: 11 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -2842,19 +2842,18 @@ SELECT count(*)
28422842
QUERY PLAN
28432843
-------------------------------------------------------------------------------
28442844
GroupAggregate
2845-
Group Key: t1.z, t1.w, t1.x, t1.y
2846-
-> Incremental Sort
2847-
Sort Key: t1.z, t1.w, t1.x, t1.y
2848-
Presorted Key: t1.z, t1.w, t1.x
2845+
Group Key: t1.x, t1.y, t1.z, t1.w
2846+
-> Sort
2847+
Sort Key: t1.x, t1.y, t1.z, t1.w
28492848
-> Merge Join
2850-
Merge Cond: ((t1.z = t2.z) AND (t1.w = t2.w) AND (t1.x = t2.x))
2849+
Merge Cond: ((t1.w = t2.w) AND (t1.z = t2.z) AND (t1.x = t2.x))
28512850
-> Sort
2852-
Sort Key: t1.z, t1.w, t1.x
2851+
Sort Key: t1.w, t1.z, t1.x
28532852
-> Index Scan using btg_x_y_idx on btg t1
28542853
-> Sort
2855-
Sort Key: t2.z, t2.w, t2.x
2854+
Sort Key: t2.w, t2.z, t2.x
28562855
-> Index Scan using btg_x_y_idx on btg t2
2857-
(13 rows)
2856+
(12 rows)
28582857

28592858
RESET enable_nestloop;
28602859
RESET enable_hashjoin;
@@ -2878,12 +2877,11 @@ SELECT count(*) FROM btg GROUP BY w, x, y, z ORDER BY x*x, z;
28782877
Sort
28792878
Sort Key: ((x * x)), z
28802879
-> GroupAggregate
2881-
Group Key: x, y, w, z
2882-
-> Incremental Sort
2883-
Sort Key: x, y, w, z
2884-
Presorted Key: x, y
2880+
Group Key: w, x, y, z
2881+
-> Sort
2882+
Sort Key: w, x, y, z
28852883
-> Index Scan using btg_x_y_idx on btg
2886-
(8 rows)
2884+
(7 rows)
28872885

28882886
-- Test the case where the number of incoming subtree path keys is more than
28892887
-- the number of grouping keys.

‎src/test/regress/expected/partition_join.out

Lines changed: 9 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -615,39 +615,39 @@ SELECT a, b FROM prt1 FULL JOIN prt2 p2(b,a,c) USING(a,b)
615615
-> Sort
616616
Sort Key: (COALESCE(prt1.a, p2.a)), (COALESCE(prt1.b, p2.b))
617617
-> Merge Full Join
618-
Merge Cond: ((prt1.a = p2.a) AND (prt1.b = p2.b))
618+
Merge Cond: ((prt1.b = p2.b) AND (prt1.a = p2.a))
619619
Filter: ((COALESCE(prt1.a, p2.a) >= 490) AND (COALESCE(prt1.a, p2.a) <= 510))
620620
-> Sort
621-
Sort Key: prt1.a, prt1.b
621+
Sort Key: prt1.b, prt1.a
622622
-> Seq Scan on prt1_p1 prt1
623623
-> Sort
624-
Sort Key: p2.a, p2.b
624+
Sort Key: p2.b, p2.a
625625
-> Seq Scan on prt2_p1 p2
626626
-> Group
627627
Group Key: (COALESCE(prt1_1.a, p2_1.a)), (COALESCE(prt1_1.b, p2_1.b))
628628
-> Sort
629629
Sort Key: (COALESCE(prt1_1.a, p2_1.a)), (COALESCE(prt1_1.b, p2_1.b))
630630
-> Merge Full Join
631-
Merge Cond: ((prt1_1.a = p2_1.a) AND (prt1_1.b = p2_1.b))
631+
Merge Cond: ((prt1_1.b = p2_1.b) AND (prt1_1.a = p2_1.a))
632632
Filter: ((COALESCE(prt1_1.a, p2_1.a) >= 490) AND (COALESCE(prt1_1.a, p2_1.a) <= 510))
633633
-> Sort
634-
Sort Key: prt1_1.a, prt1_1.b
634+
Sort Key: prt1_1.b, prt1_1.a
635635
-> Seq Scan on prt1_p2 prt1_1
636636
-> Sort
637-
Sort Key: p2_1.a, p2_1.b
637+
Sort Key: p2_1.b, p2_1.a
638638
-> Seq Scan on prt2_p2 p2_1
639639
-> Group
640640
Group Key: (COALESCE(prt1_2.a, p2_2.a)), (COALESCE(prt1_2.b, p2_2.b))
641641
-> Sort
642642
Sort Key: (COALESCE(prt1_2.a, p2_2.a)), (COALESCE(prt1_2.b, p2_2.b))
643643
-> Merge Full Join
644-
Merge Cond: ((prt1_2.a = p2_2.a) AND (prt1_2.b = p2_2.b))
644+
Merge Cond: ((prt1_2.b = p2_2.b) AND (prt1_2.a = p2_2.a))
645645
Filter: ((COALESCE(prt1_2.a, p2_2.a) >= 490) AND (COALESCE(prt1_2.a, p2_2.a) <= 510))
646646
-> Sort
647-
Sort Key: prt1_2.a, prt1_2.b
647+
Sort Key: prt1_2.b, prt1_2.a
648648
-> Seq Scan on prt1_p3 prt1_2
649649
-> Sort
650-
Sort Key: p2_2.a, p2_2.b
650+
Sort Key: p2_2.b, p2_2.a
651651
-> Seq Scan on prt2_p3 p2_2
652652
(43 rows)
653653

0 commit comments

Comments
 (0)