On Sun, 1 Jun 2008 15:36:14 -0500 "Adam Rich" <adam.r@xxxxxxxxxxxxx> wrote: > I am not familiar with the python library, but that looks correct > to me. You can always test it by adding a sleep between your two > queries and modifying the database from a console connection > during the sleep. > Note that I'm assuming your 5 queries are all read-only selects. > If you're modifying data during your queries, and another > concurrent database connection modifies the same data during your > transaction, the later modifications will fail under serializable > isolation. Which one will fail? the second query or the serializable transaction. My understanding was that the serializable transaction will fail. -- Ivan Sergio Borgonovo http://www.webthatworks.it
