RE: [python-sybase] Fix for rowcount

[<Sha...@su...>]

I believe the current behavior is correct according to
the DB API spec, because when a query is executed, but
no fetch has been performed, the interface cannot determine
the number of rows produced (second clause, not first).

    The attribute is -1 in case no executeXXX() has been performed on
    the cursor or the rowcount of the last operation is not determinable
    by the interface.[7]

Peter's problems with his patch indicate that the Sybase
module decides on its own when exactly to perform fetches
(until the time they are actually needed, of course). In
this case, the consistent behavior of returning -1 until
the user actually tries to fetch something sounds preferable.

To the best of my knowledge, Sybase libraries do not reliably
report an empty query until you actually try to fetch something,
so the problem is not with the Python module.

Peter: If you absolutely need to know the number of rows before
you try to fetch anything, you can select count(*) with the same
tables and conditions, before your query. This will always return
one row... In theory, it could be done in the same query (using
the Sybase "compute" clause, returning a separate result set),=20
but IIRC the compute results are returned only after the "regular"
results, so you could only get to the rowcount this way by first
discarding all the rows.

Hope this helps,
	Shai.