- From: Leo Sauermann <leo@gnowsis.com>
- Date: Wed, 29 Oct 2003 10:38:14 +0100
- To: "'Emmanuel Pietriga'" <epietriga@yahoo.fr>
- Cc: <www-rdf-interest@w3.org>
> > normally: > > SELECT ?x FROM (?x rdf:type foaf:Person) > > > > > > I assumed to use "." as the context node abbrevation, this is also > > standard in filesystems and XPath syntax, so commonly used. > > > I agree with this notation choice. > > > The easy syntax by me would return all "?" missing things. > > > > There could be more than one ? which could be merged > > > What do you mean by merged? (? rdf:type foaf:Person), (? rdf:type dc:Document) should return all persons and document resources in a single list. > > > > to do ANDs and ORs you could use named things: > > (?a rdf:type foaf:Person) AND (?a foaf:knowsAbout ?) > > > > how do you interpret the expression above w.r.t the context node? context is not evaluated, these search in all triples of the graph. like XPath : "//person/" > the problem is not just addressing RDFPath expressions, but also > instructions associated with these expressions. What I meant by > select="*" was not just how to have something equivalent in > rdfpath, but > what should be considered as equivalent, given the fact that > we have a > potentially cyclic graph and not a tree? no problem, Jena does show this very good: A graph consists of triples, and those are counted. All queries work on the triples. If you do a select "*" you get all triples, which is ok, if you need it and have a limited model. greetings Leo www.gnowsis.com
Received on Wednesday, 29 October 2003 04:33:09 UTC