Date: Thu, 14 Jan 2010 16:18:14 +0100
From: "A. Kretschmer" <andreas.kretschmer@schollglas.com>
To: pgsql-performance@postgresql.org
Subject: Re: Slow "Select count(*) ..." query on table with 60 Mio. rows
Message-ID: <20100114151813.GH30196@a-kretschmer.de>
References: <1263481124.3307.59.camel@zitrone>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1263481124.3307.59.camel@zitrone>
User-Agent: Mutt/1.5.9i

In response to tom :
> Hi,
> 
> === Problem ===
> 
> i have a db-table "data_measurand" with about 60000000 (60 Millions)
> rows and the following query takes about 20-30 seconds (with psql):
> 
> mydb=# select count(*) from data_measurand;
>   count   
> ----------
>  60846187
> (1 row)
> 
> 
> === Question ===
> 
> - What can i do to improve the performance for the data_measurand table?

Short answer: nothing.

Long answer: PG has to check the visibility for each record, so it
forces a seq.scan.

But you can get an estimation, ask pg_class (a system table), the column
reltuples there contains an estimated row rount.
http://www.postgresql.org/docs/current/static/catalog-pg-class.html

If you really needs the correct row-count you should create a TRIGGER
and count with this trigger all INSERTs and DELETEs.


Regards, Andreas
-- 
Andreas Kretschmer
Kontakt:  Heynitz: 035242/47150,   D1: 0160/7141639 (mehr: -> Header)
GnuPG: 0x31720C99, 1006 CCB4 A326 1D42 6431  2EB0 389D 1DC2 3172 0C99