Alan Hodgson wrote:
> It's because everything is cached, in particular the relevant rows from
> the "email" table (accessing which took 22 of the original 27 seconds).
> The plan looks good for what it's doing.
> I don't see that query getting much faster unless you could add a lot more
> cache RAM; 30K random IOs off disk is going to take a fair bit of time
> regardless of what you do.
Thanks Alan, I guessed that the caching was the difference, but I do not
understand why there is a heap scan on the email table? The query seems
to use the email_fts_index correctly, which only takes 6 seconds, why
does it then need to scan the email table?
Sorry If I sound a bit stupid - I am not very experienced with the
In response to
pgsql-performance by date
|Next:||From: Howard Cole||Date: 2008-06-17 10:54:12|
|Subject: Re: Tsearch2 Initial Search Speed|
|Previous:||From: Chris Mair||Date: 2008-06-16 19:26:14|
|Subject: Re: [pgsql-performance] function difference(geometry,geometry) is SLOW!|