Slightly different approach to rq23-like features

Some time ago I posted a compromise dataset design that offered most of
the features of rq23, but with fewer requirements on the implementation:
http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JanMar/0440.html
[its quite long an rambling, so I dont suggest you read it]

Anyway, Andy and I have been duiscussing it on IRC, and he requested a
testcase:

dataset:

	a.rdf:
	   :x :prop "1" .

	b.rdf:
	   :x :prop "2" .

	c.rdf:
	   :x :prop "3" .

	a.rdf and b.rdf are the default graph set (n.b. there is no
	anonymous graph).

query:
	SELECT ?z WHERE { ?x ?y ?z . }
result:
	-------
	| ?z  |
	-------
	| "1" |
	| "2" |
	-------

query:
	SELECT ?g ?z WHERE { GRAPH ?g { ?x ?y ?z . } }
result:
	-----------------
	| ?g      | ?z  |
	-----------------
	| <a.rdf> | "1" |
	| <b.rdf> | "2" |
	| <c.rdf> | "3" |
	-----------------

Received on Thursday, 2 June 2005 15:03:07 UTC