CODEWITHSUNDEEP
sparqldbpedia
user10867087
user10867087

Reputation: 51

DBPedia-France (fr.dpbedia.org/sparql) query giving different result from a call to another

I'm calling on http://fr.dbpedia.org/sparql the following SPARQL query:

prefix owl: <http://www.w3.org/2002/07/owl#>
prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>

select distinct ?lcs  where {

   {
     ?lcs ^(rdf:type/rdfs:subClassOf*)   <http://fr.dbpedia.org/resource/Honoré_Daumier> , 
                                         <http://fr.dbpedia.org/resource/Auguste_Rodin>;
        a owl:Class .
        filter not exists {
           ?llcs ^(rdf:type/rdfs:subClassOf*)  <http://fr.dbpedia.org/resource/Honoré_Daumier> , 
                                               <http://fr.dbpedia.org/resource/Auguste_Rodin>;
           a owl:Class ;
           rdfs:subClassOf+ ?lcs .
        }
   }
 }

On some call I've http://dbpedia.org/ontology/Person as result, on others call I'm getting http://dbpedia.org/ontology/Person and http://dbpedia.org/ontology/Agent and with others the previous answers plus http://www.w3.org/2002/07/owl#Thing

without nothing to know that a response isn't complete. How can I use the result, if it is a little randomized

Upvotes: 2

Views: 162

Answers (1)

UninformedUser
UninformedUser

Reputation: 8465

The main reason for your query not working as expected is that the data the data is i) split into separate graphs and ii) not all graphs were added to the default graph.

To keep it short, the instance data is contained inside the graph http://fr.dbpedia.org whereas the schema triples will be accessible via http://dbpedia.org graph only. Sometimes, if no graph is given, then the union of some graphs is used as the default graph which will be the dataset at query time. Unfortunately, this doesn't hold for the French DBpedia endpoint, only the instance data graph will be used.

You can check this with

DESCRIBE <http://dbpedia.org/ontology/Person>

which is empty when using either no graph explicitly or the graph http://fr.dbpedia.org, but non-empty for graph http://dbpedia.org.

The way to define the default graph is using the keyword FROM. For your query, it should therefore be 

prefix owl: <http://www.w3.org/2002/07/owl#>
prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>

select distinct ?lcs  
from <http://fr.dbpedia.org> 
from <http://dbpedia.org> 
where {
     ?lcs ^(rdf:type/rdfs:subClassOf*)   <http://fr.dbpedia.org/resource/Honoré_Daumier> , 
                                         <http://fr.dbpedia.org/resource/Auguste_Rodin>;
        a owl:Class .
        filter not exists {
           ?llcs ^(rdf:type/rdfs:subClassOf*)  <http://fr.dbpedia.org/resource/Honoré_Daumier> , 
                                               <http://fr.dbpedia.org/resource/Auguste_Rodin>;
           a owl:Class ;
           rdfs:subClassOf+ ?lcs .
        }

 }

Note, while this seems to return the correct result, you should also consider the comment from @TallTed regarding possible differences among language chapters (e.g. English vs French Wikipedia as source), release dumps (2016 vs 2018 or even the DBpedia Live) as well as Virtuoso versions used as backend.

Upvotes: 1

Related Questions