Andrew Sullivan wrote:
> On Wed, Jan 17, 2007 at 01:52:25PM +0100, Marcin St?pnicki wrote:
>> I see this in postgresql logs:
>>
>> [6210] DEBUG: query: select * from "zew_patients"
>> [6210] DEBUG: query: SELECT "A1"."p_nr_pesel" FROM "zew_patients"
>> "A1" WHERE ("A1"."p_patient_id" = 19300)
>>
>> The point is, I don't why oh why the first select is issued (some cache?),
>
> <speculation level="rank">
>
> I bet it's getting the column list from the table or some such thing.
> This is a lousy way to do it (the information_schema would be more
> correct, although maybe no faster).
That'd be my guess. And then it's not fetching any rows, expecting
cursor-like behaviour. Of course we fetch all the rows before returning
any results.
The real solution would be to add "LIMIT 0" or "LIMIT 1" to the
column-finding query, but I doubt that's possible with the Oracle plugin.
Perhaps check if there's a "fetch N rows at a time" option for the ODBC
setup that might help you.
-- Richard Huxton Archonet Ltd