Two "identical" answers to a request (utf8 and non-utf8)

63 views Asked by At

I am having problems with a live.dbpedia SPARQL request, for it returns some entries twice (once as an utf8 URI, once as a non-utf8 URI : Here are the results.

Is it something that needs to be fixed inside of dbpedia (where should it be reported)?

Is there a way to keep only one version of these duplicated urls? (I do not want to ignore a non-utf8 URI if there is no utf8 counterpart)

P.S.: The actual request

select distinct ?name where {
   ?name <http://purl.org/dc/terms/subject><http://dbpedia.org/resource/Category:Individual_graphs>.
   } ORDER BY desc(?name) LIMIT 2   
1

There are 1 answers

3
Joshua Taylor On BEST ANSWER

Even though there are multiple URIs that can identify the article, they all have the same article title, so you can extract the title (it's the value of the rdfs:label property), group by that, and then sample the URIs. Doing that, along with using the built-in DBpedia namespaces, I end up with this query:

select distinct (sample(?name_) as ?name) where {
  ?name_ dcterms:subject category:Individual_graphs ;
        rdfs:label ?label
}
group by ?label
order by desc(?name)

SPARQL results