Hi Michael,
We already tried to explain the Oracle query execution plan with our DBA.
The result is that the Oracle server prefers to make a full table scan
instead of using the index; the reason is still indeterminate (the database
statistics are up to date), we checked arguments data type and as
Hi,
Maybe you’ve run into this:
https://groups.google.com/forum/#!topic/sqlalchemy/8Xn31vBfGKU
Gabor
From: sqlalchemy@googlegroups.com [mailto:sqlalchemy@googlegroups.com] On
Behalf Of Thierry Florac
Sent: 12 May 2014 23:13
To: sqlalchemy
Subject: [sqlalchemy] Oracle index not used on
OK well of course also, as we have the exact same thing being asked in regards
to Oracle right now in another thread, you can of course always bypass a
bound value in the most direct way, using text() or literal_column():
q = s.query(Something).filter(Something.foo = literal_column('my value'))
On May 13, 2014, at 4:21 AM, Thierry Florac tflo...@gmail.com wrote:
Hi Michael,
We already tried to explain the Oracle query execution plan with our DBA. The
result is that the Oracle server prefers to make a full table scan instead of
using the index; the reason is still indeterminate
Hi Michael,
I just tried using literal_column function and now performances are as
good as they can :-/
As my problem is only limited to a single use case, using this method, even
if not perfect (arguments must be carefully verified!), if far better than
a global option.
Many thanks,
Thierry
On Mon, May 12, 2014 at 6:02 PM, Sylvester Steele sylvesterste...@gmail.com
wrote:
On Mon, May 12, 2014 at 4:53 PM, Michael Bayer
mike...@zzzcomputing.comwrote:
On May 12, 2014, at 2:30 PM, Sylvester Steele sylvesterste...@gmail.com
wrote:
the code here isn’t really showing me the
On May 13, 2014, at 2:23 PM, Sylvester Steele sylvesterste...@gmail.com wrote:
ODBC connection pooling setting did not matter. After the above change, code
is running in both cases (ODBC connection pooling on or pooling off).
Let me know if this is an issue and you need more info.
OK so
Yep, literal_column() fixed my performance problem.
Thanks again for all your help
On Tuesday, May 13, 2014 6:15:39 AM UTC-4, Michael Bayer wrote:
OK well of course also, as we have the exact same thing being asked in
regards to Oracle right now in another thread, you can of course always