Re: Workaround: Planner preference for tsquery filter vs. GIN index in fast text search

From: Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>
To: obartunov(at)gmail(dot)com, Laurence Parry <greenreaper(at)hotmail(dot)com>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Workaround: Planner preference for tsquery filter vs. GIN index in fast text search
Date: 2014-04-22 06:28:23
Message-ID: 53560C07.80207@vmware.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On 04/20/2014 07:46 AM, Oleg Bartunov wrote:
> btw, 9.4 should be wiser in case of rare+common terms, thanks to GIN
> fast scan feature.

Indeed, although we didn't actually do anything to the planner to make
it understand when fast scan helps. Doing something about cost
estimation is still on the 9.4 Open Items list, but I don't have any
ideas on what to do about it, and I haven't heard anything from
Alexander about that either. That means that the cost estimation issue
Laurence saw is going to be even worse in 9.4, because GIN is going to
be faster than a seq scan in more cases than before and the planner
doesn't know about it.

- Heikki

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Heikki Linnakangas 2014-04-22 06:40:37 Re: tsearch2, large data and indexes
Previous Message Sergey Konoplev 2014-04-22 02:01:54 Re: Best practice question