If you have a multibytes character set, the OCI API
from Oracle will for some calls count characters and
for others bytes. This will confuse DBD::Oracle especially
on CLOBs, but also on strings with limited length in
some cases.
BR, Stefan
-----Ursprüngliche Nachricht-----
Von: Sharon O'Connor [mailto:sharon(_at_)netopia(_dot_)com]
Gesendet: Donnerstag, 14. Dezember 2000 19:55
An: Ken Shan
Cc: Rami Friedman; perl-unicode(_at_)perl(_dot_)org
Betreff: Re: unicode/locale question
Yes you should have NLS_LANG=american_america.UTF8
where the language and territory are whatever the database was
originally set
up in as this is the language that messages from the database
will be returned
in.
Ken Shan wrote:
On 2000-12-14T11:55:57-0800, Rami Friedman wrote:
Could I instead rely on the database driver to convert from the
foreign charset to unicode?
I don't see why not. Have you looked at the Oracle documentation for
its NLS support? The Oracle driver should be happy to perform the
encoding conversions you require based on your setting of environment
variables and so on.
--
Edit this signature at
http://rodimus.digitas.harvard.edu/cgi-bin/ken/sig
vi has detected that it is not your default text editor.
If you would like vi to be your default text editor, please press "Y".
~
:Y
------------------------------------------------------------------------
Part 1.2Type: application/pgp-signature
--
Sincerely,
Sharon O'Connor
Netopia
(650) 966-8420 ext. 105