Horton

Re: [HACKERS] Slow count(*) again...

Joshua D. Drake <jd@commandprompt.com>

View thread
From: "Joshua D. Drake" <jd@commandprompt.com>
To: Conor Walsh <ctw@adverb.ly>
Cc: Robert Haas <robertmhaas@gmail.com>, "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>
Date: 2011-02-04T02:33:30Z
Lists: pgsql-hackers, pgsql-performance 
On Thu, 2011-02-03 at 18:12 -0800, Conor Walsh wrote:
> > I can't remember
> > anyone ever complaining "ANALYZE took too long to run".  I only
> > remember complaints of the form "I had to remember to manually run it
> > and I wish it had just happened by itself".
> 
> Robert,
> 
> This sounds like an argument in favor of an implicit ANALYZE after all
> COPY statements, and/or an implicit autoanalyze check after all
> INSERT/UPDATE statements.

Well that already happens. Assuming you insert/update or copy in a
greater amount than the threshold for the 

autovacuum_analyze_scale_factor

Then autovacuum is going to analyze on the next run. The default is .1
so it certainly doesn't take much.

JD

> 
> -Conor
> 

-- 
PostgreSQL.org Major Contributor
Command Prompt, Inc: http://www.commandprompt.com/ - 509.416.6579
Consulting, Training, Support, Custom Development, Engineering
http://twitter.com/cmdpromptinc | http://identi.ca/commandprompt