Jan Wieck <JanWieck@Yahoo.com> writes:
> The detail I'm wondering about most is how you'd know in an
> UPDATE case which two tuples (one deleted during this XACT
> and one inserted) are the two for OLD and NEW in the call to
> the trigger.
Ugh ... good point. There's no back-link from the updated tuple to
its original on disk, is there?
Back to the drawing board ...
regards, tom lane