RDF::Query#graph_name does not behave has documented about rdf HOT 3 CLOSED

I think this is a case where the documentation is out of sync with the intended implementation. The fact that a variable does not match against the default graph stems from SPARQL semantics, where querying using an unbound graph name variable will not match the default graph. See 13.3.1 Accessing Graph Names:

PREFIX foaf: <http://xmlns.com/foaf/0.1/>

SELECT ?src ?bobNick
FROM NAMED <http://example.org/foaf/aliceFoaf>
FROM NAMED <http://example.org/foaf/bobFoaf>
WHERE
  {
    GRAPH ?src
    { ?x foaf:mbox <mailto:[email protected]> .
      ?x foaf:nick ?bobNick
    }
  }

Within RDF.rb, this happens because RDF::Query::Variable#eql?(false) always returns false, which it needs to do to be able to effectively do SPARQL queries. You can see how this is explicitly defined for RDF::Repository#query_pattern:

Of course, a different Repository implementation could do this differently, but I believe it would fail various SPARQL tests.

I believe the corrective action here is to update the method signatures in RDF::Query to match this behavior. I'm afraid I don't have an easy solution to query against both the default and named graphs and get back the bound graph name. If you query a repository without using a variable for the graph_name position, you will get back all solutions, but you will not know what graph the solution came from.

from rdf.

Ok, that makes sense. Thank you for the write-up!

I guess I'm also confused by the fact that I use the SPARQL gem to build queries, and the way it handles the default graph differs from the way our usual triple store handles it.

I'll leave the issue open for that adjustment in the documentation, but other than that, it's good for me.

from rdf.

Ok, that makes sense. Thank you for the write-up!

I guess I'm also confused by the fact that I use the SPARQL gem to build queries, and the way it handles the default graph differs from the way our usual triple store handles it.

I'll leave the issue open for that adjustment in the documentation, but other than that, it's good for me.

Some triple stores treat the default graph as the union of all graphs in the dataset, but this is not a normative requirement. You can likely use the MergeGraph class from https://github.com/ruby-rdf/rdf-aggregate-repo/blob/3.3.0/lib/rdf/aggregate_repo/merge_graph.rb to simulate this behavior, but it could also be a pretty trivial wrapper around an RDF::Dataset/Repository as well.

from rdf.