- From: Hugh Glaser <hugh@glasers.org>
- Date: Wed, 10 Dec 2014 13:01:55 +0000
- To: Brent Shambaugh <brent.shambaugh@gmail.com>
- Cc: "semantic-web@w3.org" <semantic-web@w3.org>
Great question Brent. And the reminder of the SPARQL endpoint list is useful, Luca. (Although I think the uptime in the last 24 hours testing many have gone a bit awry.) I am guessing, since you talk about a "triple pattern”, not a SPARQL query, that you want to do something very simple and practical, such as: SELECT ?label WHERE {<http://dbpedia.org/resource/Stem_cell> rdfs:label ?label} That’s a classic for people to use this stuff, and there should be an easy answer. Of course, this is now a search activity, not a follow-your-nose, and so Google taught us that you need to move to central resource, and that things could still work like that. So the thing you could also do is simply lean on those nice people over at OpenLink Software and use their LDO Cloud cache. http://lod.openlinksw.com/sparql They are doing all the harvesting for us - or at least trying, and so can do better than we can on our own! Then the simply query above to their endpoint may get you what you want. But there is a possible problem - maybe the OpenLink people can comment? (Maybe in a new thread if it is likely to be a long answer.) You won’t get any sameAs stuff, as far as I can tell. So SELECT * WHERE {<http://nl.dbpedia.org/resource/Stamcel> rdfs:label ?o} doesn’t give the same results as SELECT * WHERE {<http://dbpedia.org/resource/Stem_cell> rdfs:label ?o} even though there is a owl:sameAs triple between those two subject URIs in the store. I am not an experienced Virtuoso user, so it may be that there is a well-known thing to make it happen. Best Hugh > On 10 Dec 2014, at 08:19, Luca Matteis <lmatteis@gmail.com> wrote: > > Hi Brent, > > Yes you would need an index of the whole Web unfortunately. Or at > least the portion of which you think your triple may reside in. It > could be however that your triple is being served directly by the > provider, which means you could simply try resolving the subject URI > of the triple. > > I guess instead of keeping an index you can try querying as many > SPARQL endpoints as you can for your pattern. Here's a list of > endpoints to get you started > http://labs.mondeca.com/sparqlEndpointsStatus/ > > Hope that helps > Luca > > On Wed, Dec 10, 2014 at 2:57 AM, Brent Shambaugh > <brent.shambaugh@gmail.com> wrote: >> Dear all, >> >> I am trying to figure out how it is possible to query n possible sparql >> endpoints for a particular triple pattern. >> >> I seem to think that a federated query provided by sparql only allows >> explicitly mentioned sparql endpoints. >> >> http://www.w3.org/TR/sparql11-federated-query/ >> >> It appears that powder allows filtering based on different parts of the URI: >> >> http://www.w3.org/TR/powder-grouping/ >> >> Would I need some sort of web index that catalogues known triples that I can >> then use a reference point for sparql querying? >> >> Sorry about my lack of knowledge. >> >> -Brent Shambaugh >> >> > -- Hugh Glaser 20 Portchester Rise Eastleigh SO50 4QS Mobile: +44 75 9533 4155, Home: +44 23 8061 5652
Received on Wednesday, 10 December 2014 13:02:21 UTC