Peter Geoghegan <pg@bowt.ie> writes:
> We don't need to be able to assume that binary equality is exactly the
> same thing as opclass equality at the level of individual tuples. We
> only need to be able to assume that the user cannot observe any
> differences when they are shown output for two datums that are
> opclass-equal for any opclass that supports deduplication (i.e. cases
> like the numeric_ops case just won't work, so we shouldn't event try).
Hmm, so that would exclude the optimization for numeric, float4/float8,
and nondeterministic text collations. Anything else?
I agree that teaching opclasses to say whether this is okay is a
reasonable approach.
> Consumers of this new infrastructure probably won't be limited to the
> deduplication feature;
Indeed, we run up against this sort of thing all the time in, eg, planner
optimizations. I think some sort of "equality is precise" indicator
would be really useful for a lot of things.
regards, tom lane