- From: Axel Polleres <axel.polleres@deri.org>
- Date: Wed, 6 Oct 2010 09:36:31 -0400
- To: Andy Seaborne <andy.seaborne@epimorphics.com>
- Cc: "SPARQL Working Group" <public-rdf-dawg@w3.org>
> The subquery
> SELECT-UNION-WHERE-{} can be used to insert constants but breaks the
> streaming possibilities given by BINDINGS which present query, then
> variables then the data.
Just to understand this:
What you say here is that the syntactic sugar of "BINDINGS" makes it easier to process that bit for engines that understand it than the uglier, but equivalent version with UNION subqueries, yes?
(which would coincide with my undertanding)
Axel
On 6 Oct 2010, at 04:12, Andy Seaborne wrote:
>
>
> On 05/10/10 20:13, Axel Polleres wrote:
> > looking over the minutes from last time and trying to understand the conclusion of the BINDINGS discussion...
> >
> > Do I miss anything or is semantically just the same as a Subquery:
> >
> > i.e.
> >
> > WHERE { P ]
> > BINDINGS ?X1 ... ?Xn { ( c11 ... c1n ) (c21 ... c2n) ... }
> >
> >
> > just boils down to
> >
> >
> > WHERE { { P }
> > { {SELECT ( c11 AS ?X2 ) ... ( c1n AS ?Xn ) WHERE {} }
> > UNION
> > {SELECT ( c21 AS ?X2 ) ... ( c2n AS ?Xn ) WHERE {} }
> > UNION
> > ...
> > } }
>
> Sort of. That syntax translation needs to put the constants inside
> WHERE clause and needs tweaking to cope with UNDEF. The subquery
> SELECT-UNION-WHERE-{} can be used to insert constants but breaks the
> streaming possibilities given by BINDINGS which present query, then
> variables then the data.
>
> Andy
>
Received on Wednesday, 6 October 2010 13:44:45 UTC