"Pavan Deolasee" <[EMAIL PROTECTED]> writes:
> On 9/21/07, Tom Lane <[EMAIL PROTECTED]> wrote:
>> Shouldn't we be able to prune rows that have been inserted and deleted
>> by the same transaction? I'd have hoped to see this example use only
>> one heap page ...
>>
> Not before the transaction commits ? In the test, we update a single tuple
> 10000 times in the same transaction. So there is no opportunity to prune.
[ looks a bit more ... ] Hm, the test I was thinking of was this one
at the end of HeapTupleSatisfiesVacuum:
if (TransactionIdEquals(HeapTupleHeaderGetXmin(tuple),
HeapTupleHeaderGetXmax(tuple)))
{
/*
* Inserter also deleted it, so it was never visible to anyone else.
* However, we can only remove it early if it's not an updated tuple;
* else its parent tuple is linking to it via t_ctid, and this tuple
* mustn't go away before the parent does.
*/
if (!(tuple->t_infomask & HEAP_UPDATED))
return HEAPTUPLE_DEAD;
}
but control never gets that far because neither xmin nor xmax is
committed yet. We could fix that, probably, by considering the
xmin=xmax case in the xmin-in-progress case further up; but the
HEAP_UPDATED exclusion is still a problem. Still, it seems like this
is leaving some money on the table when you think about pruning a HOT
chain. Can we improve on it easily?
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster
