Re: Allow GiST opcalsses without compress\decompres functions

From: Alexander Korotkov <a(dot)korotkov(at)postgrespro(dot)ru>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Andrey Borodin <x4mmm(at)yandex-team(dot)ru>, Dmitriy Sarafannikov <dsarafannikov(at)yandex(dot)ru>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Allow GiST opcalsses without compress\decompres functions
Date: 2017-09-20 13:31:26
Message-ID: CAPpHfdvL93RE29ud2K9FNnpG3i2Ti=pcHtvfghwgH+qUSdNNcQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Sep 20, 2017 at 4:25 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Andrey Borodin <x4mmm(at)yandex-team(dot)ru> writes:
> > You mentioned that probably there cannot be TOASTed values in the index.
> > I need to settle this question in more deterministic way.
> > Can you point where to look at the code or who to ask:
> > Can there be TOASTed values in index tuples?
>
> Yes. We don't allow out-of-line values, but we do allow compressed and
> short-header values.
>
> If you don't believe me, look at index_form_tuple().
>

OK.
So GiST opclass should either implement decompress method with detoasting
or detoast every input key in all other methods.

BTW, this comment looks still invalid for me...

> #define SIGLENINT 4 /* >122 => key will *toast*, so very slow!!! */

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Amit Kapila 2017-09-20 13:34:55 Re: Page Scan Mode in Hash Index
Previous Message Tom Lane 2017-09-20 13:25:12 Re: Allow GiST opcalsses without compress\decompres functions