My old post (http://www.jroller.com/mmatthews/entry/speeding_up_batch_inserts_for) on the performance gains from batch rewritten statements gets regurgitated in various forums, and conference sessions, and often people miss the nuances of it.
Under the hood, what is happening with this feature is the following:
(1) The driver attempts to detect that the SQL being prepared is an INSERT. We (on purpose) don‘t ship a full-fledged parser in the driver, so this works 95% of the time. For the other 5%, you‘re out of luck unless you …[Read more...]