- From: Matthew Perry <matthew.perry@oracle.com>
- Date: Mon, 25 Jul 2011 11:27:11 -0400
- To: public-rdf-dawg@w3.org
Thanks Andy. That clears it up.
- Matt
On 7/25/2011 10:48 AM, Andy Seaborne wrote:
> Matt,
>
> Any unbound variable in GRAPH becomes a lop on the URIs known to be in the dataset:
>
> eval(D(G), Graph(var,P)) =
> Let R be the empty multiset
> foreach IRI i in D
> R := Union(R, Join( eval(D(D[i]), P) , Ω(?var->i) )
> the result is R
>
> http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalGraph
>
> so it tries each one with some value of ?g making the active graph that graph "D(D[i])".
>
> eval(D(G), ZeroLengthPath(x:term, path, vy:var)) =
> { { (vy, x) } }
> card[{ (vy, x) }] = 1
>
> so it's (1), repeated or each named graph.
>
> Andy
>
> On 21/07/11 20:17, Matthew Perry wrote:
>> Hi,
>>
>> I have come across a property path query that I'm not sure about:
>>
>> SELECT ?g ?o
>> WHERE { GRAPH ?g { <urn:s> <urn:p1>{0} ?o } }
>>
>> Do we get:
>> 1) ?o=<urn:s>, ?g=<urn:graph_i> for all graphs in the dataset
>> 2) ?o=<urn:s>, ?g=<urn:graph_i> for all graphs in the dataset that
>> contain <urn:s> as a subject
>> 3) something else
>>
>> I don't believe we hit the 0-length case with an unbound graph in our
>> tests.
>>
>> Thanks,
>> Matt
>>
>
Received on Monday, 25 July 2011 15:27:46 UTC