database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2026-10-23 13:29:09 +03:00
Code Activity

Attempt to stabilize plan of self-join test in tidscan.sql

Browse Source 
The test that checks the expected plan for this self-join test has been
known to have failed in the past due to badly timed VACUUMs causing
small variations in row estimates on one of the tables, resulting in a
swapped join order.  Currently, failures have only been seen in v14, and
seemingly due to 74388a1ac and 4496020e6 the failures have not been seen
in more recent versions.

Here we shrink down the number of matching rows on one side of the join
to make the alternative join order's costs more expensive relative to
the cheapest join order.  Previously the alternative order had the same
cost.

We do this in all supported versions to reduce the chances of future
changes reintroducing stability issues with these queries.

Reported-by: Alexander Lakhin <exclusion@gmail.com>
Author: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/f5d1f4c2-6224-4797-be17-c86e77f96c9c@gmail.com
Backpatch-through: 14
This commit is contained in:
David Rowley
2026-08-21 12:27:56 +12:00
parent 7a392b0f65
commit 203231cb5a
2 changed files with 26 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
src/test/regress/expected/tidscan.out
View File

@@ -237,7 +237,8 @@ ROLLBACK;
-- (these plans don't use TID scans, but this still seems like an
-- appropriate place for these tests)
EXPLAIN (COSTS OFF)
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
QUERY PLAN
----------------------------------------
Aggregate
@@ -246,17 +247,20 @@ SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
-> Seq Scan on tenk1 t1
-> Hash
-> Seq Scan on tenk1 t2
(6 rows)
Filter: (ten = 0)
(7 rows)
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
count
-------
10000
1000
(1 row)
SET enable_hashjoin TO off;
EXPLAIN (COSTS OFF)
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
QUERY PLAN
-----------------------------------------
Aggregate
@@ -268,12 +272,14 @@ SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
-> Sort
Sort Key: t2.ctid
-> Seq Scan on tenk1 t2
(9 rows)
Filter: (ten = 0)
(10 rows)
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
count
-------
10000
1000
(1 row)
RESET enable_hashjoin;

16
src/test/regress/sql/tidscan.sql
View File

@@ -86,12 +86,20 @@ ROLLBACK;
-- (these plans don't use TID scans, but this still seems like an
-- appropriate place for these tests)
EXPLAIN (COSTS OFF)
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
SET enable_hashjoin TO off;
EXPLAIN (COSTS OFF)
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
SELECT count(*) FROM tenk1 t1 JOIN tenk1 t2 ON t1.ctid = t2.ctid
WHERE t2.ten = 0;
RESET enable_hashjoin;
-- check predicate lock on CTID