Re: Sorting Problem in UNICODE/german

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Andreas Seltenreich <andreas+pg(at)gate450(dot)dyndns(dot)org>
Cc: pgsql-bugs(at)postgresql(dot)org, Klaus Ita <postgres(at)stro(dot)at>
Subject: Re: Sorting Problem in UNICODE/german
Date: 2005-09-02 14:25:24
Message-ID: 13915.1125671124@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Andreas Seltenreich <andreas+pg(at)gate450(dot)dyndns(dot)org> writes:
>> Klaus Ita schrob:
>>> this is also wrong (There should be 'Unter' and then 'U:berflieger'
>>> [berflieger]). so is this a libc bug?

> I think I got your point now. Libc appears to be using iso-14651
> sorting for all "de" locales. I'm afraid you will have compile a
> customized locale to depart from that.

I wouldn't call it a libc bug, but a bug in the locale definition.
In any case it doesn't appear to be Postgres' problem --- if we sort
the same way "sort" does under the same locale setting, then we are
doing what we expect.

I think at this point Klaus needs to find some people who know about
hacking locale definitions. I sure don't know enough about them to
help further. Is there a libc mailing list anywhere?

One thing I do know --- if you install a new version of the locale
Postgres is using, you'd better re-initdb, or at least REINDEX all
your indexes on textual columns. Changing sort order is equivalent
to making such indexes corrupt.

regards, tom lane

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Tom Lane 2005-09-02 14:53:40 Re: BUG #1859: 3-octet private use UTF8 chars reported as identical
Previous Message Viswanath Ramineni 2005-09-02 14:13:34 BUG #1860: Insert failed due to unique index