I wanted to update those who expressed interest in Oracle support for
SQLObject and have waited so patiently (I presume). As I said a while
back, I had 0.6.1 working. Since then I've learned to run the unit
tests (I'm embarrassed to say), some of which fail, and have not found
enough time to fix these, or to do a 0.7 merge. If anyone wants to see
it in this condition, I'll post it or email it. Otherwise I hope to get
back to it in the not-to-distant future.
A few details...
There was some concern about quoting special characters, but this seems
to just work.
BLOBS don't really work right, as the DB module for Oracle, cx_Oracle,
doesn't have a binary method. I wrote code to convert between Python
and Oracle ascii hex, but this only works up to the 4,000 characters
that can be in-line in a sql statement. The proposed usage of
parameters lists should make this problem go away. (The same limitation
applies to PickleCol since it uses BLOBCol).
Oracle doesn't have a data type for a date without a time, so at this
point DateCol and DateTimeCol do the same thing (but the unit test fails).
Matthew
Matthew Cahn wrote:
> I have modified SQLObject 0.6.1 to support Oracle (except for the
> ability to create classes from an existing database, which I still
> hope to get to). Shortly I'll be merging these changes into 0.7 and
> putting them into the subversion repository.
>
> If anyone would like to try my modified 0.6.1 I'd be happy to supply it.
>
> Matthew Cahn
>
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Matthew Cahn
% Principal Systems Analyst
% Bristol-Myers Squibb Company
% Mailbox code: H23-05
% P.O. Box 4000
% Princeton, NJ 08543-4000
%
% Phone: (609) 252-3477
% Fax: (609) 252-6030
% Email: Matthew.Cahn@...
% Wireless email: matthew_cahn@...
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|