Re: Full text indexing preformance! (long)

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Mitch Vincent" <mitch(at)huntsvilleal(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Full text indexing preformance! (long)
Date: 2000-05-30 05:49:55
Message-ID: 718.959665795@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

"Mitch Vincent" <mitch(at)huntsvilleal(dot)com> writes:
> The query is very fast now (.0.039792 seconds to be exact).

Cool ...

> In my paging system I only have a need for 10 records at a time so I LIMIT
> the query. The problem comes when I need to get a total of all the records
> that matched the query (as a good search engine, I must tell people how many
> records were found).. I can't count() and LIMIT in the same query, so I'm
> forced to do 2 queries, one with count() and one without.

Well, of course the whole *point* of LIMIT is that it stops short of
scanning the whole query result. So I'm afraid you're kind of stuck
as far as the performance goes: you can't get a count() answer without
scanning the whole query.

I'm a little curious though: what is the typical count() result from
your queries? The EXPLAIN outputs you show indicate that the planner
is only expecting about one row out now, but I have no idea how close
that is to the mark. If it were really right, then there'd be no
difference in the performance of LIMIT and full queries, so I guess
it's not right; but how far off is it?

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2000-05-30 06:06:08 Re: Header File cleanup.
Previous Message Bruce Momjian 2000-05-30 02:55:02 Re: Applying TOAST to CURRENT