From: | Matthew Wakeling <matthew(at)flymine(dot)org> |
---|---|
To: | valgog <valgog(at)gmail(dot)com> |
Cc: | pgsql-performance(at)postgresql(dot)org |
Subject: | Re: Full text search with ORDER BY performance issue |
Date: | 2009-07-21 14:01:43 |
Message-ID: | alpine.DEB.2.00.0907211454180.19493@aragorn.flymine.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
On Tue, 21 Jul 2009, valgog wrote:
> There is a problem with GIN and GIST indexes, that they cannot be used
> by the ORDER BY. Maybe it will be a nice idea to ask Oleg to make it
> possible to use the b-tree columns in GIST or GIN to make the sort
> easier, but I have no idea how difficult it will be to implement it in
> current GIN or GIST structures. I think Oleg or even Tom will be the
> right people to ask it :)
I can answer that one for GiST, having had a good look at GiST recently.
There is simply no useful information about order in a GiST index for it
to be used by an ORDER BY. The index structure is just too general,
because it needs to cope with the situation where a particular object type
does not have a well defined order, or where the "order" is unuseful for
indexing.
Matthew
--
A good programmer is one who looks both ways before crossing a one-way street.
Considering the quality and quantity of one-way streets in Cambridge, it
should be no surprise that there are so many good programmers there.
From | Date | Subject | |
---|---|---|---|
Next Message | Kevin Grittner | 2009-07-21 14:02:02 | Re: Calling conventions |
Previous Message | Grzegorz Jaśkiewicz | 2009-07-21 13:53:44 | Re: hyperthreaded cpu still an issue in 8.4? |