From: Tom Lane <tgl@sss.pgh.pa.us>
To: David Gould <daveg@sonic.net>
cc: Alexander Kuzmenkov <a.kuzmenkov@postgrespro.ru>,
        pgsql-hackers@lists.postgresql.org,
        Alina Alexeeva <alexeeva@adobe.com>,
        Ullas Lakkur Raghavendra <lakkurra@adobe.com>
Subject: Re: [patch] BUG #15005: ANALYZE can make pg_class.reltuples
 inaccurate.
In-reply-to: <20180313001440.581b6662@engels>
References: <20180117164916.3fdcf2e9@engels>
 <fa8122c2-ea7a-ff38-49ac-e6616aec38ea@postgrespro.ru>
 <20180228192337.67cc3f99@engels>
 <4988c54a-1709-a978-9a43-65f8a1e09e94@postgrespro.ru>
 <18522.1519916974@sss.pgh.pa.us>
 <1386effe-0e6d-fd5a-6f0d-f294e4ef2dda@postgrespro.ru>
 <3240.1519948160@sss.pgh.pa.us> <20180302161840.3530921f@engels>
 <28773.1520357803@sss.pgh.pa.us> <6981.1520871694@sss.pgh.pa.us>
 <20180313001440.581b6662@engels>
Comments: In-reply-to David Gould <daveg@sonic.net>
	message dated "Tue, 13 Mar 2018 00:14:40 -0700"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <996.1520954439.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Tue, 13 Mar 2018 11:20:39 -0400
Message-ID: <997.1520954439@sss.pgh.pa.us>
Precedence: bulk

David Gould <daveg@sonic.net> writes:
> I also thought about the theory and am confident that there really is no=
 way
> to trick it. Basically if there are enough pages that are different to a=
ffect
> the overall density, say 10% empty or so, there is no way a random sampl=
e
> larger than a few hundred probes can miss them no matter how big the tab=
le is.
> If there are few enough pages to "hide" from the sample, then they are s=
o few
> they don't matter anyway.

> After all this my vote is for back patching too. I don't see any case wh=
ere
> the patched analyze is or could be worse than what we are doing. I'm hap=
py to
> provide my test cases if anyone is interested.

Yeah, you have a point.  I'm still worried about unexpected side-effects,
but it seems like overall this is very unlikely to hurt anyone.  I'll
back-patch (minus the removal of the unneeded vac_estimate_reltuples
argument).

			regards, tom lane