- From: Axel Polleres <axel.polleres@deri.org>
- Date: Wed, 2 Feb 2011 16:29:39 +0000
- To: Andy Seaborne <andy.seaborne@epimorphics.com>
- Cc: SPARQL Working Group <public-rdf-dawg@w3.org>
Looks ok to me, I like the purchase order example, it's a very illlustrative example for multiple paths, thanks! Axel On 30 Jan 2011, at 19:02, Andy Seaborne wrote: > http://www.w3.org/2009/sparql/wiki/CommentResponse:JP-3 > > Drafted - ready for review > > > (I've sent a version of this to Jorge offlist as discussions aren't > possible on the comments list). > > So he's seen the gist of the reply, > > Andy > > On 26/01/11 14:39, Andy Seaborne wrote: > > > > > > On 26/01/11 13:53, Matthew Perry wrote: > >> I think this should work for the sum age query. > >> > >> SELECT SUM(?A) > >> WHERE > >> { ?F :age ?A > >> { SELECT DISTINCT ?F > >> WHERE > >> { :me (:friend)+ ?F } }} > >> > >> It seems that, in general, DISTINCT will work as a there exists query > >> when URIs and blank nodes are the endpoint of a path because we have a > >> one-to-one mapping between URI/BN values and graph vertices, but we get > >> a one-to-many mapping from literal values to vertices. > > > > Yes, that example addresses Jorge's use case and contrasts nicely with > > the basket example where there is a need for duplicate matches from the > > property path itself (although, in the syntax triple patterns are > > property paths of length 1). > > > > Andy > >
Received on Wednesday, 2 February 2011 16:30:16 UTC