Re: [SQLObject] PostgreSQL: costly count

[Frank B. <fb...@fo...>]

Hallo,
J-P Lee hat gesagt: // J-P Lee wrote:

> Hmm.. the thread seems to suggest the same thing.  I'm not sure if you 
> can get around the sequential scan.  count(id) saves you the time of 
> fetching all field contents.  

Fetching field contents is not the problem, I guess, because a count
doesn't (or shouldn't) fetch any contents (we didn't ask for them).

This message however:
http://archives.postgresql.org/pgsql-performance/2003-10/msg00093.php
suggests doing a "count(pkey)" instead: 
  
  <quote>
  Have you tried doing
  
  SELECT count(pkey)
  
  rather than count(*) where pkey is the primary key (assuming you
  have a single field that is a primary key or a unique indexed key).
  This is MUCH faster in my experience.  
  </qoute>

Note that the pkey is not the id-colmun, but an index on this column.

I couldn't try this yet, but it might be worth a try.

ciao
-- 
 Frank Barknecht                               _ ______footils.org__