Re: Wich hardware suits best for large full-text indexed

Oleg Bartunov wrote:

>Diogo,
>
>could you send us 'explain analyze' for fts-specific part of your
>query ? For example,
>
>
>
Alright, here it goes:

eyeknow=# explain analyze select * from documentos where
conteudo_stem_ix @@ to_tsquery('default_portuguese', 'brasil');

QUERY PLAN

-----------------------------------------------------------------------------------------------------------------------------------------------------------
 Index Scan using documentos_conteudo_stem_ix_ix on documentos
(cost=0.00..1183.08 rows=291 width=1156) (actual time=26.33..840039.40
rows=70686 loops=1)
   Index Cond: (conteudo_stem_ix @@ '\'brasil\''::tsquery)
   Filter: (conteudo_stem_ix @@ '\'brasil\''::tsquery)
 Total runtime: 840248.91 msec
(4 rows)

>Frankly, I don't understand your query :) Could you explain what do you
>want to find ?
>
>qrydocumentos.conteudo_stem_ix @@ to_tsquery('default_portuguese', 'brasil')
>
>
I'm trying to find documents which contain the word 'brasil' (stemmed).
This is just an example, because the user can input any query to the
database.
The rest of the query is used to make sure that the user is accessing
only documents he has permission to.

Regards,

--
Diogo Biazus
diogo@ikono.com.br
http://www.ikono.com.br