Re: EARL results for RDFLib against data-r2/graph/manifest.ttl

Chimezie Ogbuji wrote:
> Attached is a test report documenting an RDFLib run against the
> data-r2/graph/manifest.ttl tests.  It passes all the
> data-r2/graph/graph*.rq tests.  I followed the testing convention we are
> using in the GRDDL WG of including DOAP RDF statements about the test
> subject (RDFLib in this case) as well as FOAF RDF statements about the
> tester (myself).  I've attached both Turtle and RDF/XML serializations
> of the test run.  There was one test that seemed malformed (below).
> 
> ## graph-02 ##
> 
> SELECT * { ?s ?p ?o }
> 
> There are two triples in the named graph
> (http://www.w3.org/2001/sw/DataAccess/tests/data-r2/graph/data-g1.ttl)
> and there is *no* default graph.  Is this dataset well-formed? Section 8
> (RDF Dataset) says: "an RDF Dataset *always* contains one default
> graph").

The default graph always exists - it's a question of what it is loaded with.

SELECT ...
FROM <a>
FROM <b>

puts two graphs into the default graph.

Similarly, in the test manifest, if there are two qt:data then two graphs are 
read into the default graph, not two default graphs.

That said, the idea is not to test the test manifest (:-) so we could put in 
an empty graph for clarity.

 Andy

> 
> So, it is not immediately clear to me what the BGP is matching against
> (i.e., what is the 'active' graph), especially when 8.3 (Querying the
> Dataset) says: "Outside the use of GRAPH, the default graph is matched
> by basic graph patterns." - however, there is no default graph in this
> dataset.  Unless the interpretation is that the default graph is an
> empty graph.
> 
> 

-- 
Hewlett-Packard Limited
Registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Monday, 9 July 2007 14:56:37 UTC