Re: [PATCH] Don't block HOT update by BRIN index - Mailing list pgsql-hackers

From Josef Šimánek
Subject Re: [PATCH] Don't block HOT update by BRIN index
Date
Msg-id CAFp7QwqFsKucYsftGVMAuFWTG5-tGy6EFYAfMzOj=UzZLsAuZA@mail.gmail.com
Whole thread Raw
In response to Re: [PATCH] Don't block HOT update by BRIN index  (Tomas Vondra <tomas.vondra@enterprisedb.com>)
Responses Re: [PATCH] Don't block HOT update by BRIN index
List pgsql-hackers
st 30. 6. 2021 v 1:20 odesílatel Tomas Vondra
<tomas.vondra@enterprisedb.com> napsal:
>
>
>
> On 6/30/21 12:53 AM, Josef Šimánek wrote:
> > st 30. 6. 2021 v 0:31 odesílatel Josef Šimánek <josef.simanek@gmail.com> napsal:
> >>
> >> Hello!
> >>
> >> Tomáš Vondra has shared a few ideas to improve BRIN index in czech
> >> PostgreSQL mail list some time ago [1 , in czech only]. This is first
> >> try to implement one of those ideas.
> >>
> >> Currently BRIN index blocks HOT update even it is not linked tuples
> >> directly. I'm attaching the initial patch allowing HOT update even on
> >> BRIN indexed columns. This patch went through an initial review on
> >> czech PostgreSQL mail list [1].
> >
> > I just found out current patch is breaking partial-index isolation
> > test. I'm looking into this problem.
> >
>
> The problem is in RelationGetIndexAttrBitmap - the existing code first
> walks indnatts, and builds the indexattrs / hotblockingattrs. But then
> it also inspects expressions and the predicate (by pull_varattnos), and
> the patch fails to do that for hotblockingattrs. Which is why it fails
> for partial-index, because that uses an index with a predicate.
>
> So there needs to be something like:
>
>      if (indexDesc->rd_indam->amhotblocking)
>          pull_varattnos(indexExpressions, 1, &hotblockingattrs);
>
>      if (indexDesc->rd_indam->amhotblocking)
>          pull_varattnos(indexPredicate, 1, &hotblockingattrs);
>
> This fixes the failure for me.

Thanks for the hint. I'm attaching a fixed standalone patch.

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

Attachment

pgsql-hackers by date:

Previous
From: Tomas Vondra
Date:
Subject: Re: [PATCH] Don't block HOT update by BRIN index
Next
From: Alvaro Herrera
Date:
Subject: Re: Preventing abort() and exit() calls in libpq