- From: Olaf Hartig <hartig@informatik.hu-berlin.de>
- Date: Fri, 4 Jun 2010 14:37:16 +0200
- To: public-rdf-dawg-comments@w3.org
- Message-Id: <201006041437.16670.hartig@informatik.hu-berlin.de>
Hello, The example in Section 2.1 of the Federation Extensions document [1] uses the void vocabulary wrongly. The default graph in this example contains RDF triples such as <http://GABABR1.example/SPARQL> a void:Dataset ; dcterms:subject entrez:h2550 . The variable ?service of the example query is then mapped to URI <http://GABABR1.example/SPARQL> and this URI is then used to identify a SPARQL service. That's not the meaning of void:Dataset. void:Dataset refers to a linked dataset and not to a SPARQL endpoint that exposes a linked dataset. I propose to adjust the example as follows: 1.) replace the abovementioned triples by the following (similar for the other SPARQL services): <http://GABABR1.example/dataset> a void:Dataset ; dcterms:subject entrez:h2550 ; void:sparqEndpoint <http://GABABR1.example/SPARQL> . 2.) replace the first triple patterns in the sample query by the following: # Find the service with the expertise. ?dataset dcterms:subject ?gene ; void:sparqEndpoint ?service . FILTER (?gene = entrez:h2550 || ?gene = entrez:h9568) BTW, the FILTER clause in the current example lacks a closing bracket. Greetings, Olaf [1] http://www.w3.org/TR/2010/WD-sparql11-federated-query-20100601/
Received on Friday, 4 June 2010 12:38:50 UTC