Jan Wieck <JanWieck@Yahoo.com> writes:
> Aah - there is the first bullet hole in my multi-ctid-index-idea. Now
> the question becomes how expensive these tests are (if a normal backend
> can do them at all within reason)?
It's not hugely expensive, IIRC, you just need to make some additional
checks against global xmin (compare HeapTupleSatisfiesVacuum against
the others). We're already doing something similar for the optimization
that suppresses subsequent heap lookups for globally-dead index tuples.
I'm dubious about the multi-ctid idea though because it would mean
bloating the index tuple header, whether there was any use for multiple
entries or not. (Could we make the header variable size? Not sure it's
worth the trouble.)
regards, tom lane