I converted data to RDF in our research institute. I use triple stores to search the RDF data such as Virtuoso, OWLIM-SE (GraphDB) and Fuseki.
I'd like to connect with many external databases on the fly by SPARQL endpoint. However I couldn't use more than one "SERVICE" query. Here is the sample query.
prefix b2r: <http://bio2rdf.org/>
prefix drgb: <http://bio2rdf.org/drugbank_vocabulary/>
prefix dbowl: <http://dbpedia.org/ontology/>
prefix orphan: <http://www.nibio.go.jp/orphanDrugTarget#>
select distinct ?title ?protein_name
where {
?s dbowl:drugbank ?drugbank.
SERVICE <http://drugbank.bio2rdf.org/sparql>
{?drugbank <http://purl.org/dc/terms/title> ?title;
<http://bio2rdf.org/drugbank_vocabulary:target> ?target.
?target <http://bio2rdf.org/drugbank_vocabulary:x-uniprot> ?x_uniprot.
BIND (SUBSTR(str(?x_uniprot),28) as ?uniprot_id)
BIND (URI(CONCAT("http://purl.uniprot.org/uniprot/",?uniprot_id)) as ?uniprot)}
SERVICE <http://beta.sparql.uniprot.org/>
{?uniprot <http://purl.uniprot.org/core/mnemonic> ?protein_name}}
Although I tried the query 3 triple stores, I failed to get results. I'd already tested one "SERVICE" query could work and these queries could work individual.
If you know how to solve the problem (I'd like to avoid that all data put into our triple store.), please help me.
Thank you for your reply. (This post is almost the same for BioHackathon ML, sorry) I installed Virtuoso 7. If it is caused by bug, it's a pity,
Kato-san told me in case of virtuoso, the change of settings have possibilities to solve the problem. http://answers.semanticweb.com/questions/10360/virtuoso-federated-query
Although I could overcome an error, I couldn't get any results...
It seems this Q and A (semanticweb.com) includes some hints.
http://answers.semanticweb.com/search/?csrfmiddlewaretoken=599f1bddf88920d231c9236be45453cd&q=federated+query&Submit=search&t=question
Still investigating...
We use OWLIM-SE too, I'd like to get more information about OWLIM-SE too.