[DB-SIG] Re: [Python-Dev] Toward Python's future article

Chris Cogdon chris at cogdon.org
Mon Oct 11 19:21:01 CEST 2004


On Oct 11, 2004, at 09:41, M.-A. Lemburg wrote:

> I'm not sure I follow you here: the database modules usually do
> document which object types they use on output and recognize on
> input.

Its all very well to document it, but it means that differing database 
connectors can still return different types, even when staying on the 
same back-end database. In other words, this is still 'module-centric' 
rather than 'application-centric'.

Yes, I agree that database with very strange types (such as 
PostgreSQL's line and shape types), the module-writer is going to have 
to come up with unique python types, but things like 'date' are as 
stable as 'int'.

I DO like the idea of feeding handlers to the database, so the 
application writer can decide what kind of types to return in what 
circumstances.


And... I'd like to point out that if more application-writer-centric 
features are put into DB-API 3.0, it does not preclude a module writer 
continuing to write to DB-API 2.0 instead (and, perhaps, letting 
someone else write a 3.0 wrapper)



-- 
    ("`-/")_.-'"``-._        Chris Cogdon <chris at cogdon.org>
     . . `; -._    )-;-,_`)
    (v_,)'  _  )`-.\  ``-'
   _.- _..-_/ / ((.'
((,.-'   ((,/   fL



More information about the DB-SIG mailing list