Problem with Oracle interfaces and non-ASCII characters

[Dietmar Schwertberger]

Hi!

I'm accessing an Oracle database via cx_Oracle, DCOracle2 and odbc from the
win32 extensions.
Unfortunately both native Oracle interfaces don't return non-ASCII
characters.
E.g. I'm getting a '?' instead of the degree symbol '\xb0' or 'A' instead of
the A umlaut '\xc4' (in ISO latin 1 encoding).
The odbc module returns these characters as intended, but I'd like to avoid
my users having to set up ODBC first.

The network traffic with the native interfaces contains the non-ASCII
characters, so the conversion to ASCII happens in either the oracle oci.dll
or the Python interfaces.

The database is version 8i, but I'm not sure about that; client is 8.1.7
and an older version (7.x) for ODBC.
I can't test the native oracle interfaces with the 7.x client as I don't
have binaries for that (Python 2.2 or 2.1/Zope).

Any suggestions?


Regards,

Dietmar