To: Jamie Lawrence <jal@jal.org>
Cc: pgsql-sql@postgresql.org
Subject: Re: Different encodings in different DBs in same cluster 
In-reply-to: <20050721183224.GA19811@clueinc.net> 
References: <20050721183224.GA19811@clueinc.net>
Comments: In-reply-to Jamie Lawrence <jal@jal.org>
	message dated "Thu, 21 Jul 2005 14:32:24 -0400"
Date: Sun, 24 Jul 2005 14:41:44 -0400
Message-ID: <21993.1122230504@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

Jamie Lawrence <jal@jal.org> writes:
> I see:

>    Since these locale settings are frozen by initdb, the apparent
>    flexibility to use different encodings in different databases of a
>    cluster is more theoretical than real.

> Does anyone know what "more theoretical than real" mean in this context?

It means there are some locales that actively fail (you get inconsistent
comparison and sorting behavior) when presented with multibyte data that
doesn't match their encoding expectations.  IMHO such locale definitions
are broken and should be fixed, but they are not under our control.

> If I set the locale to C, is it going to work correctly with UTF8
> encoded data?

C will work "correctly" for suitably small values of "correctly" ---
non-ASCII characters may not sort where you'd wish, and it won't know
anything about case-folding for non-ASCII characters.  But it will at
least give consistent results.

When you use a non-C locale, it's best to stick to the encoding that
the locale expects.

			regards, tom lane