Re: Tsearch2 performance on big database

Oleg Bartunov wrote:
>
> stat() is indeed a bigdog, it was designed for developers needs,
> so we recommend to save results in table.
>
>>
>> Anyway, here's my pg_ts_cfgmap now (well the relevant bits):
>>
>> default_english | lhword       | {en_ispell,en_stem}
>> default_english | lpart_hword  | {en_ispell,en_stem}
>> default_english | lword        | {en_ispell,en_stem}
>>
>> Is it normal that queries for single words (or perhaps they are words
>> that are common) take a really long time? Like this:
>>
>
> 'hispanic' isn't common, I see you get only  674 rows and 'buckingham &
> palace'  returns 185 rows. Did you run 'vacuum analyze' ?
> I see a big discrepancy between estimated rows (8041) and actual rows.
>
>

Yes, I did a vacuum analyze right before executing these queries.

I'm going to recreate the gist index now, and do a vacuum full analyze
after that.. see if that makes a difference.

Rick

--
Systems Administrator for Rockingstone IT
http://www.rockingstone.com
http://www.megabooksearch.com - Search many book listing sites at once