Re: Detoast iterators -take 2

From: Aleksander Alekseev <aleksander(at)tigerdata(dot)com>
To: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Cc: Nikita Malakhov <hukutoc(at)gmail(dot)com>
Subject: Re: Detoast iterators -take 2
Date: 2025-08-26 13:10:54
Message-ID: CAJ7c6TP8xjcRUFxX2qKW+hjAhok=UC_6oSCV7BW47rar6H+k6g@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Nikita,

Thanks for working on this.

> master:
> [...]
> Time: 47.346 ms
>
> patched:
> [...]
> Time: 7.607 ms

Looks impressive. I believe there is still some room for improvement though:

1. IMO the new API (create_detoast_iterator et al) should have
corresponding unit tests.
2. Suggestion: perhaps (entity)_(action) naming would be better, that
is: detoast_iterator_create(), detoast_iterator_free() etc.
3. If you could demonstrate that there is no performance degradation
for other scenarios that would be great.

--
Best regards,
Aleksander Alekseev

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Antonin Houska 2025-08-26 13:31:33 Re: Adding REPACK [concurrently]
Previous Message Tomas Vondra 2025-08-26 12:56:44 Re: index prefetching