On Tue, 2 Apr 2002, Tom Lane wrote:
> Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:
> > Are the array iterator functions supposed to act sort of like
> > =ANY/=ALL except across an array instead of a subselect?
>
> Seems like a reasonable definition.
>
> > If so,
> > isStrict probably isn't right, since for an empty subselect the return
> > value does not depend on the element being searched for.
>
> Hm ... isn't it NULL anyway, if the left side is NULL?
>
> But if you're right, then the correct fix involves updating the
> functions to V1 calling conventions, so that they can make a correct
> test for NULL inputs (rather than bogusly checking for zero value).
I think the empty case is special, due to the rules on <quantified
comparison predicate>s. It looks like no comparison predicates
need to be run.
I think the applicable parts of 8.7 are
General Rule 2a
If T is empty or if the implied <comparison predicate> is true
for every row RT in T, then "R <comp op> <all> T" is true.
General Rule 2d
If T is empty or if the implied <comparison predicate> is false
for every row RT in T, then "R <comp op> <some> T" is false.