On Mon, Mar 28, 2022 at 5:21 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> In any case, what I was on about is _bt_preprocess_keys() and
> adjacent code. I'm surprised that those aren't more expensive
> than one palloc in _bt_first. Maybe that logic falls through very
> quickly in simple cases, though.
I assume that it doesn't really appear in very simple cases (also
common cases). But delaying the scan setup work until execution time
does seem ugly. That's probably a good enough reason to refactor.
--
Peter Geoghegan