Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> Using the same page for an updated tuple is generally a useful
> optimization, so I don't think you have much hopes for having it
> disabled.
Especially not since there's no very reasonable way for anything as
low-level as heap_update to know that "the table is way beyond its
nominal size". What's nominal size anyway?
Actually, the recent thinking in this area has been to try to *increase*
the usage of same-page UPDATE, so as to prevent table bloat in the first
place ...
regards, tom lane