Re: Query performance question

On Wed, 28 Mar 2001, David Olbersen wrote:

> There are two things you should do first:
>
> 1) VACUUM ANALYZE hits;
> 2) VACUUM ANALYZE referrer;

I did that before, too.

> If it's still running slowly, try EXPLAINing your SELECT to us.

Ok, VACUUMing the tables didn't increase performance. This is the EXPLAIN
output:

httplog=# EXPLAIN SELECT count(*), url FROM hits, referrer WHERE
referrer.id = referrer_id GROUP BY url ORDER BY count DESC LIMIT 10;
NOTICE:  QUERY PLAN:

Sort  (cost=912.81..912.81 rows=678 width=20)
  ->  Aggregate  (cost=846.98..880.90 rows=678 width=20)
        ->  Group  (cost=846.98..863.94 rows=6785 width=20)
              ->  Sort  (cost=846.98..846.98 rows=6785 width=20)
                    ->  Hash Join  (cost=10.46..415.17 rows=6785 width=20)
                          ->  Seq Scan on hits  (cost=0.00..128.85
rows=6785 width=4)
                          ->  Hash  (cost=9.57..9.57 rows=357 width=16)
                                ->  Seq Scan on referrer  (cost=0.00..9.57
rows=357 width=16)

EXPLAIN

I hope someone can explain this to me :)

Thanks for your quick answer,

regards,
le

--
Lukas Ertl                          eMail: l.ertl@univie.ac.at
WWW-Redaktion                       Tel.:  (+43 1) 4277-14073
Zentraler Informatikdienst (ZID)    Fax.:  (+43 1) 4277-9140
der Universität Wien