Horton

Re: document the need to analyze partitioned tables

Laurenz Albe <laurenz.albe@cybertec.at>

View thread View on postgr.es
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Bruce Momjian <bruce@momjian.us>
Cc: Justin Pryzby <pryzby@telsasoft.com>, Nathan Bossart <nathandbossart@gmail.com>, Tomas Vondra <tomas.vondra@enterprisedb.com>, Robert Haas <robertmhaas@gmail.com>, Álvaro Herrera <alvherre@alvh.no-ip.org>, yuzuko <yuzukohosoya@gmail.com>, pgsql-hackers@postgresql.org
Date: 2023-01-20T09:33:57Z
Lists: pgsql-hackers 
On Thu, 2023-01-19 at 15:56 -0500, Bruce Momjian wrote:
> On Thu, Jan 19, 2023 at 01:50:05PM +0100, Laurenz Albe wrote:
> > On Wed, 2023-01-18 at 16:23 -0500, Bruce Momjian wrote:
> > > Is it possible to document when partition table statistics helps?
> > 
> > I think it would be difficult to come up with an exhaustive list.
> 
> I was afraid of that.  I asked only because most people assume
> autovacuum handles _all_ statistics needs, but this case is not handled.
> Do people even have any statistics maintenance process anymore, and if
> not, how would they know they need to run a manual ANALYZE?

Probably not.  I think this would warrant an entry in the TODO list:
"make autovacuum collect startistics for partitioned tables".

Even if we cannot give better advice than "run ALANYZE manually if
the execution plan looks fishy", the patch is still an improvement,
isn't it?

I have already seen several questions by people who read the current
documentation and were worried that autovacuum wouldn't clean up their
partitioned tables.

Yours,
Laurenz Albe

Commits

  1. doc: Fix typo in ANALYZE documentation

  2. Document autoanalyze limitations for partitioned tables

  3. Revert analyze support for partitioned tables

  4. Describe (auto-)analyze behavior for partitioned tables