From: | "Joshua D(dot) Drake" <jd(at)commandprompt(dot)com> |
---|---|
To: | Josh Berkus <josh(at)agliodbs(dot)com> |
Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Alvaro Herrera <alvherre(at)commandprompt(dot)com>, Greg Smith <greg(at)2ndquadrant(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: default_statistics_target WAS: max_wal_senders must die |
Date: | 2010-10-21 02:32:21 |
Message-ID: | 1287628341.7085.90.camel@jd-desktop |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Wed, 2010-10-20 at 15:15 -0700, Josh Berkus wrote:
> >> Maybe what should be done about this is to have separate sizes for the
> >> MCV list and the histogram, where the MCV list is automatically sized
> >> during ANALYZE.
>
> It's been suggested multiple times that we should base our sample size
> on a % of the table, or at least offer that as an option. I've pointed
> out (with math, which Simon wrote a prototype for) that doing
> block-based sampling instead of random-row sampling would allow us to
> collect, say, 2% of a very large table without more I/O than we're doing
> now.
>
> Nathan Boley has also shown that we could get tremendously better
> estimates without additional sampling if our statistics collector
> recognized common patterns such as normal, linear and geometric
> distributions. Right now our whole stats system assumes a completely
> random distribution.
>
> So, I think we could easily be quite a bit smarter than just increasing
> the MCV. Although that might be a nice start.
References would be nice.
JD
>
> --
> -- Josh Berkus
> PostgreSQL Experts Inc.
> http://www.pgexperts.com
>
--
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
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2010-10-21 02:45:07 | Re: default_statistics_target WAS: max_wal_senders must die |
Previous Message | Nathan Boley | 2010-10-21 02:10:39 | Re: default_statistics_target WAS: max_wal_senders must die |