BUG #19709: Incorrect result when comparing OLD tuple column to itself in INSERT RETURNING

From: PG Bug reporting form <noreply(at)postgresql(dot)org>
To: pgsql-bugs(at)lists(dot)postgresql(dot)org
Cc: 1482694023(at)qq(dot)com
Subject: BUG #19709: Incorrect result when comparing OLD tuple column to itself in INSERT RETURNING
Date: 2026-09-20 12:28:14
Message-ID: 19709-388df6818f7ef822@postgresql.org
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

The following bug has been logged on the website:

Bug reference: 19709
Logged by: N J
Email address: 1482694023(at)qq(dot)com
PostgreSQL version: 18.4
Operating system: Windows 11 64-bit
Description:

Environment:
PostgreSQL: 18.4
OS: Windows 11 64-bit

Reproduction query:

CREATE TEMP TABLE IF NOT EXISTS a0 (x INTEGER);
TRUNCATE TABLE a0;
INSERT INTO a0 VALUES (42)
RETURNING
CASE
WHEN (old).x IS DISTINCT FROM old.x THEN 1
ELSE 0
END AS bug;

Observed result:
bug
-----
1

Expected result:
bug
-----
0

Explanation:
For INSERT statements, the OLD tuple is all NULL.
The expression old.x IS DISTINCT FROM old.x compares NULL with NULL.
Per SQL standard, NULL IS DISTINCT FROM NULL returns false, so the CASE
expression should return 0.
PostgreSQL incorrectly evaluates this predicate to true and returns 1.

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Ayush Tiwari 2026-09-20 12:58:02 Re: BUG #19709: Incorrect result when comparing OLD tuple column to itself in INSERT RETURNING
Previous Message PG Bug reporting form 2026-09-20 11:57:37 BUG #19708: Hash Join becomes about 300x slower with higher work_mem