- From: Kendall Clark <kendall@monkeyfist.com>
- Date: Wed, 23 Mar 2005 11:12:37 -0500
- To: "Seaborne, Andy" <andy.seaborne@hp.com>
- Cc: DAWG Mailing List <public-rdf-dawg@w3.org>
On Wed, Mar 23, 2005 at 03:55:55PM +0000, Seaborne, Andy wrote:
> pattern-group.
>
> { pattern } UNION {}
>
> is "pattern OR true".
Ah, hmm. I think an example query like that should be in
rq23. FWIW. I'll tweak my stuff, though.
> >> | element dataset-constraint {
> >> (uri | variable), (group-pattern)
> >> }*
> >> ^^^^ ????? why *
> >
> >
> >I assume trang set that cardinality to * because of sparqlx-23.xml in
> >the zip file, which serializes this query:
>
> I'd do that as a group with two GRAPHs in it rather than special casing it.
But that wouldn't change the cardinality of dataset-constraint, would
it? You'd still have something like:
<query-pattern>
<group-pattern>
<dataset-constraint>
<uri>...</uri>
<triple-pattern>...</triple-pattern>
</dataset-constraint>
<dataset-constraint>
<variable name="foo"/>
<triple-pattern></triple-pattern>
</dataset-constraint>
</group-pattern>
</query-pattern>
Plus, I don't understand this as an interpretation of rq23 syntax:
GRAPH data:aliceFOAF { ...several triple patterns... }
Doesn't the { and } introduce a group-pattern?
Kendall
Received on Wednesday, 23 March 2005 16:16:32 UTC