From: Richard C. <ri...@cy...> - 2006-04-03 14:54:39
|
Hi Onno, On 3 Apr 2006, at 13:59, Onno Paap wrote: > I would like to combine different ontology's resident in triple > stores on different servers. Eventually, thousands of them. > I am not certain but I suppose the SPARQL syntax provides for that. Well ... SPARQL has syntax for querying multiple *graphs*. A graph can be anything with a URI. The question is, what kind of URIs can you use in the FROM clause? That depends totally on the triple store/toolkit you're using. In most SPARQL stores, you can put URIs of RDF files in the FROM clause. This means you can query multiple RDF *files* that reside on different servers. What you want to do is query multiple *endpoints*. I don't know of *any* tool that can do this yet, but it would certainly be very useful and cool. So if you build it, be sure to tell us about it! Richard > If not, I can work around it in plain old PHP programming, issuing > separate SPARQL queries for separate servers, but I would like to > know if that is the course of action or not. > > For example, I have two different SPARQL endpoints (they really > exist): > http://www.15926.org/2006/02/oim and > http://www.15926.org/2006/02/rdl > with named graphs in them of the same names. > They are not, but suppose they are on different servers. > Now I want to combine them like this: > > prefix dm: <http://www.15926.org/2006/02/dm#> > select ?id ?superclass > from < http://www.15926.org/2006/02/oim> > from <http://www.15926.org/2006/02/rdl> > where { > ?id rdfs:subClassOf ?x . > ?x owl:hasValue ?superclass . > ?superclass rdf:type dm:ClassOfInanimatePhysicalObject . > } > > Note: in RAP the namespaces rdfs, owl, and rdf don't need to be > declared with prefix. > > The top two triples in the where clause work for the .../rdl graph > and the bottom for the .../oim graph. > This combined query doesn't work in RAP SPARQL (I'm using netapi). > My question is: should it work? Or would it work with different > syntax? Or totally not? > Onno Paap |