"Ed L." wrote:
> If the ordering I seek exists via the algorithm I've described, I don't
The right ordering is commit order, because that is the order that
played back serialized (without possible lock conflicts) guarantees the
same tuples hit.
Unfortunately I know of no sure way to tell the commit order. The
max(seqid) per transaction in ascending order would be a good
approximation at least.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck@Yahoo.com #