database/postgres
database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-05-02 11:44:50 +03:00
Code

Force parallelism in partition_aggregate

Browse Source 
Commit db0d67db2 tweaked sort costing, which however resulted in a
couple plan changes in our regression tests. Most of the new plans were
fine, but partition_aggregate were meant to test parallel plans and the
new plans were serial.

Fix that by lowering parallel_setup_cost to 0, which is enough to switch
to the parallel plan again.

Report and patch by David Rowley.

Author: David Rowley
Reviewed-by: Tomas Vondra
Discussion: https://postgr.es/m/CAApHDvpVFgWzXdtUQkjyOPhNrNvumRi_=ftgS79KeAZ92tnHKQ@mail.gmail.com
This commit is contained in:
Tomas Vondra 2022-09-04 23:37:42 +02:00
parent 4209126655
commit 01474f5698
2 changed files with 63 additions and 54 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
src/test/regress/expected/partition_aggregate.out
View File

@ -942,6 +942,7 @@ INSERT INTO pagg_tab_ml SELECT i % 30, i % 10, to_char(i % 4, 'FM0000') FROM gen
ANALYZE pagg_tab_ml;
-- For Parallel Append
SET max_parallel_workers_per_gather TO 2;
SET parallel_setup_cost = 0;
-- Full aggregation at level 1 as GROUP BY clause matches with PARTITION KEY
-- for level 1 only. For subpartitions, GROUP BY clause does not match with
-- PARTITION KEY, but still we do not see a partial aggregation as array_agg()
@ -949,24 +950,18 @@ SET max_parallel_workers_per_gather TO 2;
EXPLAIN (COSTS OFF)
SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HAVING avg(b) < 3 ORDER BY 1, 2, 3;
QUERY PLAN
--------------------------------------------------------------------------------------
Sort
--------------------------------------------------------------------------------------------
Gather Merge
Workers Planned: 2
-> Sort
Sort Key: pagg_tab_ml.a, (sum(pagg_tab_ml.b)), (array_agg(DISTINCT pagg_tab_ml.c))
-> Append
-> Parallel Append
-> GroupAggregate
Group Key: pagg_tab_ml.a
Filter: (avg(pagg_tab_ml.b) < '3'::numeric)
-> Sort
Sort Key: pagg_tab_ml.a
-> Seq Scan on pagg_tab_ml_p1 pagg_tab_ml
-> GroupAggregate
Group Key: pagg_tab_ml_2.a
Filter: (avg(pagg_tab_ml_2.b) < '3'::numeric)
-> Sort
Sort Key: pagg_tab_ml_2.a
-> Append
-> Seq Scan on pagg_tab_ml_p2_s1 pagg_tab_ml_2
-> Seq Scan on pagg_tab_ml_p2_s2 pagg_tab_ml_3
-> GroupAggregate
Group Key: pagg_tab_ml_5.a
Filter: (avg(pagg_tab_ml_5.b) < '3'::numeric)
@ -975,7 +970,15 @@ SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HA
-> Append
-> Seq Scan on pagg_tab_ml_p3_s1 pagg_tab_ml_5
-> Seq Scan on pagg_tab_ml_p3_s2 pagg_tab_ml_6
(25 rows)
-> GroupAggregate
Group Key: pagg_tab_ml_2.a
Filter: (avg(pagg_tab_ml_2.b) < '3'::numeric)
-> Sort
Sort Key: pagg_tab_ml_2.a
-> Append
-> Seq Scan on pagg_tab_ml_p2_s1 pagg_tab_ml_2
-> Seq Scan on pagg_tab_ml_p2_s2 pagg_tab_ml_3
(27 rows)
SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HAVING avg(b) < 3 ORDER BY 1, 2, 3;
a | sum | array_agg | count
@ -995,22 +998,16 @@ SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HA
EXPLAIN (COSTS OFF)
SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HAVING avg(b) < 3;
QUERY PLAN
---------------------------------------------------------------------
Append
---------------------------------------------------------------------------
Gather
Workers Planned: 2
-> Parallel Append
-> GroupAggregate
Group Key: pagg_tab_ml.a
Filter: (avg(pagg_tab_ml.b) < '3'::numeric)
-> Sort
Sort Key: pagg_tab_ml.a
-> Seq Scan on pagg_tab_ml_p1 pagg_tab_ml
-> GroupAggregate
Group Key: pagg_tab_ml_2.a
Filter: (avg(pagg_tab_ml_2.b) < '3'::numeric)
-> Sort
Sort Key: pagg_tab_ml_2.a
-> Append
-> Seq Scan on pagg_tab_ml_p2_s1 pagg_tab_ml_2
-> Seq Scan on pagg_tab_ml_p2_s2 pagg_tab_ml_3
-> GroupAggregate
Group Key: pagg_tab_ml_5.a
Filter: (avg(pagg_tab_ml_5.b) < '3'::numeric)
@ -1019,8 +1016,17 @@ SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HA
-> Append
-> Seq Scan on pagg_tab_ml_p3_s1 pagg_tab_ml_5
-> Seq Scan on pagg_tab_ml_p3_s2 pagg_tab_ml_6
(23 rows)
-> GroupAggregate
Group Key: pagg_tab_ml_2.a
Filter: (avg(pagg_tab_ml_2.b) < '3'::numeric)
-> Sort
Sort Key: pagg_tab_ml_2.a
-> Append
-> Seq Scan on pagg_tab_ml_p2_s1 pagg_tab_ml_2
-> Seq Scan on pagg_tab_ml_p2_s2 pagg_tab_ml_3
(25 rows)
RESET parallel_setup_cost;
-- Full aggregation at level 1 as GROUP BY clause matches with PARTITION KEY
-- for level 1 only. For subpartitions, GROUP BY clause does not match with
-- PARTITION KEY, thus we will have a partial aggregation for them.

3
src/test/regress/sql/partition_aggregate.sql
View File

@ -222,6 +222,7 @@ ANALYZE pagg_tab_ml;
-- For Parallel Append
SET max_parallel_workers_per_gather TO 2;
SET parallel_setup_cost = 0;
-- Full aggregation at level 1 as GROUP BY clause matches with PARTITION KEY
-- for level 1 only. For subpartitions, GROUP BY clause does not match with
@ -235,6 +236,8 @@ SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HA
EXPLAIN (COSTS OFF)
SELECT a, sum(b), array_agg(distinct c), count(*) FROM pagg_tab_ml GROUP BY a HAVING avg(b) < 3;
RESET parallel_setup_cost;
-- Full aggregation at level 1 as GROUP BY clause matches with PARTITION KEY
-- for level 1 only. For subpartitions, GROUP BY clause does not match with
-- PARTITION KEY, thus we will have a partial aggregation for them.