From: | "Guillaume Smet" <guillaume(dot)smet(at)gmail(dot)com> |
---|---|
To: | "Alexander Staubo" <alex(at)purefiction(dot)net> |
Cc: | pgsql-performance <pgsql-performance(at)postgresql(dot)org> |
Subject: | Re: Estimation problem with a LIKE clause containing a / |
Date: | 2007-11-07 13:38:04 |
Message-ID: | 1d4e0c10711070538mf773791ke5950849299b142c@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers pgsql-performance |
Alexander,
On 11/7/07, Alexander Staubo <alex(at)purefiction(dot)net> wrote:
> That's a difference of less than *three milliseconds* -- a difference
> probably way within the expected overhead of running "explain
> analyze". Furthermore, all three queries use the same basic plan: a
> sequential scan with a filter. At any rate you're microbenchmarking in
> a way that is not useful to real-world queries. In what way are these
> timings a problem?
If you read my previous email carefully, you'll see they aren't a
problem: the problem is the estimation, not the timing. This is a self
contained test case of a far more complex query which uses a bad plan
containing a nested loop due to the bad estimate.
> Now all "like 'prefix%'" queries should use the index.
Not when you retrieve 50% of this table of 22k rows but that's not my
problem anyway. A seqscan is perfectly fine in this case.
Thanks anyway.
--
Guillaume
From | Date | Subject | |
---|---|---|---|
Next Message | Bruce Momjian | 2007-11-07 13:48:36 | Re: Segmentation fault using digest from pg_crypto |
Previous Message | Bruce Momjian | 2007-11-07 13:35:31 | Re: A small rant about coding style for backend functions |
From | Date | Subject | |
---|---|---|---|
Next Message | Kevin Grittner | 2007-11-07 16:32:15 | Re: index stat |
Previous Message | Alexander Staubo | 2007-11-07 13:25:40 | Re: Estimation problem with a LIKE clause containing a / |