Missing update of all_hasnulls in BRIN opclasses

From: Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com>
To: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Missing update of all_hasnulls in BRIN opclasses
Date: 2022-10-21 15:23:45
Message-ID: 402430e4-7d9d-6cf1-09ef-464d80afff3b@enterprisedb.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

While working on some BRIN code, I discovered a bug in handling NULL
values - when inserting a non-NULL value into a NULL-only range, we
reset the all_nulls flag but don't update the has_nulls flag. And
because of that we then fail to return the range for IS NULL ranges.

Reproducing this is trivial:

create table t (a int);
create index on t using brin (a);
insert into t values (null);
insert into t values (1);

set enable_seqscan = off;
select * from t where a is null;

This should return 1 row, but actually it returns no rows.

Attached is a patch fixing this by properly updating the has_nulls flag.

I reproduced this all the way back to 9.5, so it's a long-standing bug.
It's interesting no one noticed / reported it so far, it doesn't seem
like a particularly rare corner case.

regards

--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Attachment Content-Type Size
brin-hasnulls-fix.patch text/x-patch 4.6 KB

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bharath Rupireddy 2022-10-21 15:32:04 Re: Avoid memory leaks during base backups
Previous Message Bharath Rupireddy 2022-10-21 15:07:18 Re: Crash after a call to pg_backup_start()