Hi,
I wish to restrict this URI Burner query to WordNet named graph only. Is the SPARQL syntax correct (it was generated by FCT)? If the syntax is correct, why are there items in the result set that do not logically belong to WordNet named graph? This FCT article states:
graph="graph_iri"
installed default is to search in all graphs, but system defaults may override this
Is this behavior implemented at the SPARQL processor or at the FCT service endpoint? Also, is it the case that Virtuoso’s system default is overriding the graph set in the SPARQL, so that input:default-graph-uri
has no effect?
This seems to affect only the list-count
view type. The Categories and Properties seem to filter properly to WordNet Classes and Properties. The diff between the view@type=properties
query (left) and view@type=list-count
query (right) shows that the select distinct ?g
part is the only real difference, but when I remove that wrapper, the results still seem out-of-scope:
For the uriburner.com query:
define input:default-graph-uri <http://www.ontologyportal.org/WordNet#>
select ?s1 as ?c1 count (*) as ?c2
where
{ select distinct ?s1 ?g
{ quad map virtrdf:DefaultQuadMap
{ graph ?g
{ ?s1 <http://www.w3.org/2000/01/rdf-schema#label> ?o1 .
?o1 bif:contains '(POLAR AND BEAR)' .
} } } }
group by ?s1
order by desc 2
limit 30
offset 0
if you remove the graph ?g
specifier, i.e. –
define input:default-graph-uri <http://www.ontologyportal.org/WordNet#>
select ?s1 as ?c1 count (*) as ?c2
where
{ select distinct ?s1
{ quad map virtrdf:DefaultQuadMap
{ ?s1 <http://www.w3.org/2000/01/rdf-schema#label> ?o1 .
?o1 bif:contains '(POLAR AND BEAR)' .
} } }
group by ?s1
order by desc 2
limit 30
offset 0
but then the http://www.ontologyportal.org/WordNet# on uriburner.com is empty thus no results are returned, query …
Thank you Hugh, this is indeed the case.
FYI, we discovered that the proxy was submitting all request to LOD, regardless of the domain of the target data server, due to a bug in configuration external proxy library. This issue has been fixed. Thank you for helping us detect this.
@hwilliams
Here are two examples where I attempt to restrict the named graph, query 1 and query 2. As you can see, 1) the results are the same for both, and 2) there are superfluous items in each result set.
This is also tested by choosing a “Library” in the PoC UI, for instance.