Date: Tue, 8 May 2007 11:04:27 -0400
From: Bill Moran <wmoran@collaborativefusion.com>
To: "Alexander Staubo" <alex@purefiction.net>
Cc: pgsql-performance@postgresql.org
Subject: Re: specific query (not all) on Pg8 MUCH slower than Pg7
Message-Id: <20070508110427.c714d52f.wmoran@collaborativefusion.com>
In-Reply-To: <88daf38c0705080759o2e5771c3qdf1b34baa581fc53@mail.gmail.com>
References: <200705081418.l48EIY8V026360@larrys.harvard.edu>
	<1160.1178635705@sss.pgh.pa.us>
	<88daf38c0705080759o2e5771c3qdf1b34baa581fc53@mail.gmail.com>
Organization: Collaborative Fusion
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

In response to "Alexander Staubo" <alex@purefiction.net>:

> On 5/8/07, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> > You're not getting the indexscan optimization of the LIKE clause, which
> > is most likely due to having initdb'd the 8.1 installation in something
> > other than C locale.  You can either redo the initdb in C locale (which
> > might be a good move to fix other inconsistencies from the 7.3 behavior
> > you're used to) or create a varchar_pattern_ops index on the column(s)
> > you're using LIKE with.
> 
> Given the performance implications of setting the wrong locale, and
> the high probability of accidentally doing this (I run my shells with
> LANG=en_US.UTF-8, so all my databases have inherited this locale), why
> is there no support for changing the database locale after the fact?
> 
> # alter database test set lc_collate = 'C';
> ERROR:  parameter "lc_collate" cannot be changed

How would that command handle UTF data that could not be converted to C?

-- 
Bill Moran
Collaborative Fusion Inc.
http://people.collaborativefusion.com/~wmoran/

wmoran@collaborativefusion.com
Phone: 412-422-3463x4023