On Thu, 2007-06-28 at 18:17 -0400, Christopher Browne wrote:
> Are there more items we should try to add?

Is there a good way to make it more efficient to do large
deletes/updates?

For instance, if the number of tuples that need to be deleted for a
transaction exceeds a certain amount, could we use a different process
for the delete on the subscriber so that it doesn't do millions of
single-tuple deletes?

I don't know exactly how that would work, perhaps by using a temporary
table on the subscriber and doing a single "DELETE FROM foo WHERE id IN
(SELECT id FROM foo_delete_tmp)" or something?

Regards,
	Jeff Davis