"Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov> writes:
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> ItemPointerSetInvalid(&(tuple.t_self));
> Aren't those tuples pointing to a disk buffer?
Oh, I should have looked at the code before commenting ;-).
Yeah, the correct TID value would be ItemPointerGetBlockNumber(tid)
plus the current offnum. However we don't have enough information
in this function to set t_tableOid correctly, so maybe it would be
best to just set both fields invalid. Or do nothing --- AFAICS none
of the uses of the heapTuple look at those fields. Is it worth a few
extra cycles to initialize unused fields of a short-lived heapTuple?
regards, tom lane