MIME-Version: 1.0
References: <CAE2r8H7+eACA+QxisuPY1jveA=YLuH7CE8F3TedmZfsKdYxOeA@mail.gmail.com>
 <CABV9wwNYQ6V4-kKzPbJ0fGtgrq_+0z1kLiWQ8yqqDL-hEmphdQ@mail.gmail.com>
In-Reply-To: <CABV9wwNYQ6V4-kKzPbJ0fGtgrq_+0z1kLiWQ8yqqDL-hEmphdQ@mail.gmail.com>
From: VASUKI M <vasukianand0119@gmail.com>
Date: Wed, 21 Jan 2026 11:59:24 +0530
Message-ID: <CAE2r8H7jVmsLaWi19ZuxmwR-8UjtN6Ejv7YmtDrPinCYtE8cUA@mail.gmail.com>
Subject: Re: Optional skipping of unchanged relations during ANALYZE?
To: Robert Treat <rob@xzilla.net>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, ilya.evdokimov@tantorlabs.com, 
	Christoph Berg <myon@debian.org>, Robert Haas <robertmhaas@gmail.com>
Content-Type: multipart/alternative; boundary="000000000000f5d3930648e00a8a"
Archived-At: <https://www.postgresql.org/message-id/CAE2r8H7jVmsLaWi19ZuxmwR-8UjtN6Ejv7YmtDrPinCYtE8cUA%40mail.gmail.com>
Precedence: bulk

--000000000000f5d3930648e00a8a
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi all,

Thanks a lot for the thoughtful feedback.

The points about statistics targets and random sampling make sense. In
particular, I agree that even without data changes, re-running ANALYZE
can still be beneficial (for example after increasing
default_statistics_target or per-column targets, or simply to obtain a
different sample).

Given that, my intention is to keep this strictly as an opt-in
behavior, so that existing semantics are unchanged unless the user
explicitly requests it. In the current prototype, tables that have
never been analyzed before are not skipped, and SMART only considers
relations that already have statistics.

Regarding statistics targets, since PostgreSQL does not currently track
which target was used to collect existing statistics, SMART ANALYZE
would not attempt to account for target changes. I plan to document
this limitation clearly so users understand the trade-off when opting
into this mode [for now later will look into it].

I=E2=80=99ll take this feedback into account while cleaning up the patch an=
d
documentation, and will follow up with a v1 proposal once ready.

Thanks again for the guidance.

Regards,
Vasuki M
C-DAC,Chennai

--000000000000f5d3930648e00a8a
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr">Hi all,<br><br>Thanks a lot for the thoug=
htful feedback.<br><br>The points about statistics targets and random sampl=
ing make sense. In<br>particular, I agree that even without data changes, r=
e-running ANALYZE<br>can still be beneficial (for example after increasing<=
br>default_statistics_target or per-column targets, or simply to obtain a<b=
r>different sample).<br><br>Given that, my intention is to keep this strict=
ly as an opt-in<br>behavior, so that existing semantics are unchanged unles=
s the user<br>explicitly requests it. In the current prototype, tables that=
 have<br>never been analyzed before are not skipped, and SMART only conside=
rs<br>relations that already have statistics.<br><br>Regarding statistics t=
argets, since PostgreSQL does not currently track<br>which target was used =
to collect existing statistics, SMART ANALYZE<br>would not attempt to accou=
nt for target changes. I plan to document<br>this limitation clearly so use=
rs understand the trade-off when opting<br>into this mode [for now later wi=
ll look into it].<br><br>I=E2=80=99ll take this feedback into account while=
 cleaning up the patch and<br>documentation, and will follow up with a v1 p=
roposal once ready.<br><br>Thanks again for the guidance.<br><br>Regards,<b=
r>Vasuki M<br>C-DAC,Chennai</div></div>

--000000000000f5d3930648e00a8a--