And the recipe I have used is to issue a flush() after the deletes and before the inserts. In most cases this is sufficient to get things to work in the right order. I can imagine that there are some complex data management use cases where that is not sufficient. It works for your sample as the comment in your code indicates. This should not change the performance of the app since we are only changing the sequence of statements, not introducing additional transaction overhead.
When I run into scenarios where the sequence of issuing SQL is significant, I'm not sure how much control I want to give up to the UOW code. After all if sequence is important this could very well be application dependent and the UOW might have to get really sophisticated to guess right. That increases the risk of taking a performance hit in the 90+% of cases where it doesn't matter. -- Mike Conley -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com. To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en.
