Le Mardi 16 Novembre 2004 17:06, Joshua D. Drake a écrit :
> > QUERY PLAN
> >--------------------------------------------------------------------------
> >-------------------------------------- Seq Scan on site_rss s
> > (cost=0.00..11863.16 rows=295 width=158) (actual time=17.414..791.937
> > rows=12 loops=1)
> > Filter: (site_name ~~* '%atari%'::text)
> > SubPlan
> > -> Seq Scan on user_choice u (cost=0.00..3.46 rows=1 width=4)
> > (actual time=0.222..0.222 rows=0 loops=12)
> > Filter: ((id_site = $0) AND (id_user = 1))
> > Total runtime: 792.099 ms
> >
> >First time I run the request I have a result in about 789 miliseconds
> > !!???
> >
> >I'm using PostgreSQL v7.4.6 with a Bi-Penitum III 933 Mhz and 1 Gb of RAM.
> >
> >Any idea ... ? For the moment I'm going back to use the ilike solution ...
> > but I was really thinking that Tsearch2 could be a better solution ...
>
> Well I would be curious about what happens the second time you run the
> query.
> The first time is kind of a bad example because it has to push the index
> into ram.
The second time is really quicker yes ... about 312 miliseconds ...
But for each search I have after it take about 3 or 4 seconds ...
So what can I do ?
Regards,
--
Hervé Piedvache
Elma Ingénierie Informatique
6 rue du Faubourg Saint-Honoré
F-75008 - Paris - France
Pho. 33-144949901
Fax. 33-144949902