From: Stefano O. <sed...@ho...> - 2012-06-22 16:41:22
|
Hello Ivan, I can't use the query you suggested me because i need to keep the mapping between resource and type of the resource (basically i have a list of dbpedia resources and for each of them i want to retrieve all the rdf:types, doing that using only one query). I thing i'll solve the problem installing dbpedia datasets locally, so i have a question: can you tell me what find i've to download to retrieve all the informations about type? (I want the query "select ?x where <resource> rdf:type ?x" made locally returned me exactly the same results the one of dpedia endpoint)? I've already downloaded the Ontology Infobox Types from core datasets, is it enough? Cheers, Stefano On 21/giu/2012, at 19.17, Ivan Mikhailov wrote: > Hello Stefano, > > On Thu, 2012-06-21 at 18:39 +0100, Stefano Ortona wrote: >> my query is very simple: >> >> " >> select distinct ?0 ?1 ...?x ...?100 >> WHERE >> { >> { <certain-dbpedia-resource> rdf:type ?x.} >> } >> " >> >> Basically i wanna retrieve the rdf:types for 100 different dbpedia resources, is that ~200K of text? (seems strange) >> I'm using the DBPedia SPARQL endpoint, http://dbpedia.org/sparql. > > 7 lines of SQL text per triple pattern if default graph is not > specified, 9 lines if it is specified. So it's some about 700 or 900 > lines, not 10000 . Terrible but not fatal. > > I'd query > SELECT DISTINCT ?t > from some graph if it is known > WHERE > { ?s a ?t . FILTER ( ?s IN (<resource1>, <resource2>, > <resource3> ...)) } > > A single triple pattern with IN operator would be much cheaper. > > Best Regards, > > Ivan Mikhailov > OpenLink Software > http://virtuoso.openlinksw.com > > > |