[sc34wg3] TMQL: Predicate invocation and scope

[Robert Barta]

On Tue, Jul 08, 2008 at 05:36:09PM +0200, Lars Heuer wrote:
> Hi Lars,
> 
> [ $a = member-of(member: $p, group: $g)]
> > I know this doesn't work now. It's not even syntactically valid, as
> > far as I know. But we could define this in such a way that it *would*
> > work.
> 
> I don't know if it is worth, because we'd change the meaning of
> predicate invocations seriously.

Predicate invocations generate association items, so in

  select $p, $q
  where
     member-of(member: $p, group: $g)

the 'member-of(member: $p, group: $g)' gets converted (I'm simplifying)
into

    // member-of [ . -> member == $p ] [ . -> group == $g ] ..

> Not sure if it is a good or bad thing.

Not sure what you want to say here.

> What happens to the variables $p and $g if the predicate invocation
> returns associations?

I cannot see a problem. $p and $g will be instantiated with items, so
that the 'member-of(member: $p, group: $g)' gives some non-empty
result (in other words: "exists"). If you want all of it, then

  select $p, $q, $a
  where
     $a == member-of(member: $p, group: $g)

will do the obvious.

\rho