Re: POC: GROUP BY optimization

From: Alena Rybakina <a(dot)rybakina(at)postgrespro(dot)ru>
To: Andrei Lepikhov <a(dot)lepikhov(at)postgrespro(dot)ru>, Richard Guo <guofenglinux(at)gmail(dot)com>, Alexander Korotkov <aekorotkov(at)gmail(dot)com>
Cc: Pavel Borisov <pashkin(dot)elfe(at)gmail(dot)com>, vignesh C <vignesh21(at)gmail(dot)com>, PostgreSQL Developers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com>, Teodor Sigaev <teodor(at)sigaev(dot)ru>, David Rowley <dgrowleyml(at)gmail(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Subject: Re: POC: GROUP BY optimization
Date: 2024-01-15 13:21:05
Message-ID: df09eb50-b4ae-4662-9247-73296410e8dc@postgrespro.ru
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 15.01.2024 12:46, Andrei Lepikhov wrote:
> On 15/1/2024 13:42, Richard Guo wrote:
>>
>> On Mon, Jan 15, 2024 at 8:20 AM Alexander Korotkov
>> <aekorotkov(at)gmail(dot)com <mailto:aekorotkov(at)gmail(dot)com>> wrote:
>>
>>     Thank you for providing the test case relevant for this code change.
>>     The revised patch incorporating this change is attached. Now the
>>     patchset looks good to me.  I'm going to push it if there are no
>>     objections.
>>
>>
>> Seems I'm late for the party.  Can we hold for several more days?  I'd
>> like to have a review on this patch.
> Get on board! It looks like this feature needs as much review as
> possible (likewise SJE).
>
Hi! Thank you for your work on this issue! I believe that this will help
the scheduler to make a more optimal query plan here and therefore speed
up their execution.

I have reviewed patches and noticed that we can add some code
refactoring. I have attached a diff file (group_by.diff) to this email.

The changes involve spelling corrections, renaming variables and porting
some common parts.

In addition, I have a few questions, since some points in the code
remained unclear to me.

1.  I didn't understand why we have a question in the comment next to
the enable_group_by_reordering variable in
src/backend/optimizer/path/pathkeys.c file, I assumed it was spelling
and fixed it in the diff file.

2. Why do we set the variable (path = path_save) here
(add_paths_to_grouping_rel function) if we change its variable below and
we can pass path_save as a parameter?

foreach(lc2, pathkey_orderings)
{
    PathKeyInfo *info = (PathKeyInfo *) lfirst(lc2);

    /* restore the path (we replace it in the loop) */
    path = path_save;

    path = make_ordered_path(root,
                             grouped_rel,
                             path,
                             cheapest_path,
                             info->pathkeys);
    if (path == NULL)
        continue;

--
Regards,
Alena Rybakina
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

Attachment Content-Type Size
group_by.diff text/x-patch 4.1 KB

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message David Geier 2024-01-15 13:24:38 Re: postgres_fdw fails to see that array type belongs to extension
Previous Message Daniel Gustafsson 2024-01-15 12:43:03 Re: Fix a possible socket leak at Windows (src/backend/port/win32/socket.c)