On 6/26/14, 12:24 PM, Mike Bayer wrote: > > I'm not sure what kind of application this is but I would not be mixing > transactional control, that is the commit/rollback, inside of a business > method that only seeks to create some new objects. There'd be a > containing pattern within which store_enos_rec() is called these 20K > times, then the commit() is called on the outside of that, or at least > every 1000 records or so. > > Also, the "create if not present" pattern is often much better set up by > preloading all the objects you'll deal with up front: > > existing_evs = dict(Session.query(EnosValues.enos_id, EnosValues)) > if ev_id not in existing_evs: > Session.add(EnosValues(...)) > > > If you truly need to optimize a method whose job it is to start a new > transaction, insert a single row if not present and then commit the > transaction fully, I'd use a core insert() construct for that. > Spinning up a whole unit of work in that context is overkill if > performance is a concern. this is something that's pretty common and i may have the opportunity soon to come up with a helper for this case.
> > > -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To unsubscribe from this group and stop receiving emails from it, send an email to sqlalchemy+unsubscr...@googlegroups.com. To post to this group, send email to sqlalchemy@googlegroups.com. Visit this group at http://groups.google.com/group/sqlalchemy. For more options, visit https://groups.google.com/d/optout.
