To: "Kevin Grittner" <Kevin.Grittner@wicourts.gov>
cc: pgsql-performance@postgresql.org
Subject: Re: Good News re count(*) in 8.1 
In-reply-to: <43FC4384.EE98.0025.0@wicourts.gov> 
References: <43FC4384.EE98.0025.0@wicourts.gov>
Comments: In-reply-to "Kevin Grittner" <Kevin.Grittner@wicourts.gov>
	message dated "Wed, 22 Feb 2006 10:57:08 -0600"
Date: Wed, 22 Feb 2006 12:42:02 -0500
Message-ID: <13377.1140630122@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

"Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
> We are replicating data from 72 source databases, each with the
> official copy of a subset of the data, to four identical consolidated
> databases, spread to separate locations, to serve our web site and other
> organization-wide needs.  Currently, two of these central databases are
> running a commercial product and two are running PostgreSQL.  There have
> been several times that I have run a SELECT COUNT(*) on an entire table
> on all central machines.  On identical hardware, with identical data,
> and equivalent query loads, the PostgreSQL databases have responded with
> a count in 50% to 70% of the time of the commercial product, in spite of
> the fact that the commercial product does a scan of a non-clustered
> index while PostgreSQL scans the data pages.

Interesting.  I think though that the people who are complaining come
from databases where COUNT(*) takes constant time because the DB keeps
a running count in the table's metadata.

			regards, tom lane