Horton

Re: Missing update of all_hasnulls in BRIN opclasses

Matthias van de Meent <boekewurm+postgres@gmail.com>

View thread View on postgr.es
From: Matthias van de Meent <boekewurm+postgres@gmail.com>
To: Tomas Vondra <tomas.vondra@enterprisedb.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-10-21T15:50:32Z
Lists: pgsql-hackers 
On Fri, 21 Oct 2022 at 17:24, Tomas Vondra
<tomas.vondra@enterprisedb.com> wrote:
>
> 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.

Ah, that's bad.

One question though: doesn't (shouldn't?) column->bv_allnulls already
imply column->bv_hasnulls? The column has nulls if all of the values
are null, right? Or is the description of the field deceptive, and
does bv_hasnulls actually mean "has nulls bitmap"?

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

One comment on the patch:

> +SET enable_seqscan = off;
> + [...]
> +SET enable_seqscan = off;

Looks like duplicated SETs. Should that last one be RESET instead?

Apart from that, this patch looks good.

- Matthias

Commits

  1. Show empty BRIN ranges in brin_page_items

  2. Fix handling of empty ranges and NULLs in BRIN

  3. Fix handling of NULLs when merging BRIN summaries

  4. Move IS [NOT] NULL handling from BRIN support functions