From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "Carlo Stonebanks" <stonec(dot)register(at)sympatico(dot)ca> |
Cc: | pgsql-performance(at)postgresql(dot)org |
Subject: | Re: How planner decides left-anchored LIKE can use index |
Date: | 2007-09-07 00:28:16 |
Message-ID: | 6716.1189124896@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
"Carlo Stonebanks" <stonec(dot)register(at)sympatico(dot)ca> writes:
> There is an index on lower(last_name). I have seen the planner convert the
> LIKE to lower(last_name) >= 'smith' and lower(last_name) < 'smiti' on 8.2.4
> systems, but a slow sequence scan and filter on 8.1.9 - is this related to
> the version difference (8.1.9 vs 8.2.4) or is this related to something like
> operators/classes that have been installed?
Most likely you used C locale for the 8.2.4 installation and some other
locale for the other one.
In non-C locale you can still get the optimization if you create an
index using the text_pattern_ops opclass ... but beware that this index
is useless for the normal locale-aware operators.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Greg Smith | 2007-09-07 04:26:23 | Re: SAN vs Internal Disks |
Previous Message | Carlo Stonebanks | 2007-09-07 00:06:44 | How planner decides left-anchored LIKE can use index |