Re: Character codification

There is an encoding which is used inside the server. Then the client can
specify what encoding the strings are he is sending. The server then
transforms the clientencoding to the internal representation. When the
described problem occurs this means that something in this process is not set
or wrong set. In the documentation you can find a list with supported
encodings and conversions in chapter 20 of the manual.

For me it did work best to take UTF8 as server encoding and specify the
correct client encoding within the initialisation of the database connection
(if required at all since most clients are UTF8 anyway).

If you need to fix the data already in the database you might use recode or
iconv on a text backup dump and restore the result.

Best regards
Ivo Rossacher

Am Donnerstag, 21. Dezember 2006 15.30 schrieb Sistemas C.M.P.:
> I'm using Linux server with postgres 8.0.2 and windows clients with VB
> aplications that using pgOleDB.dll and Libpq.dll. The problem is:
> Some user updates a record (with Text type field) with '?' character in
> this field Then, when someone select this record the '?' characters appears
> like '$1' This only happen with a few users. Not all
> Thanks.