I found a solution: I re-run initdb with Unicode encoding and now seems to work fine.
I cannot figure out why did this happen, or does the initdb encoding affects all the databases or the db encoding is not enough to be specified?
Andy.
----- Original Message -----
Sent: Tuesday, April 25, 2006 3:51 PM
Subject: [ADMIN] Database LATIN1 to UFT-8 (unicode) converting problem
Hi to all,
We just tried to switch out database from LATIN 1 (ISO 8859-1) to UTF-8. After a few tests we found out some problems:
- the upper and lower functions does not work for all characters. Especially lower returns nothing if a umlaut is in the converted string. The umlauts are not converted at all.
- the returned rows are not correctly sorted. ö is before a and so on.
The server is a Suse 10 system. DB is Postgresql 8.0.3.
This sorting problem we had it before and it was the problem was at local system settings. Now everything is switched to UTF-8 (or it seems so).
Please advise.
Thank you,
Andy.
