See https://jena.apache.org/documentation/tdb/datasets.html#special-graph-names

You may need to use <urn:x-arq:UnionGraph> (not sure, not a feature I ever use 
myself) though not sure if tdbquery allows for overriding the default graph

Rob

On 22/01/2020, 09:58, "Élie Roux" <[email protected]> wrote:

    Dear all,
    
    I'm trying to run tdbquery on my triple store which has a UnionGraph
    (assembler attached), but it seems that even the most simple query
    doesn't work:
    
    SELECT ?subject ?predicate ?object
    WHERE {
      ?subject ?predicate ?object
    }
    LIMIT 25
    
    Looking at the results, it seems that tdbquery is running a
    quadpattern in <urn:x-arq:DefaultGraphNode>, which makes the query
    return no result:
    
    $ tdbquery --explain --time
    --loc=/usr/local/fuseki/base/databases/newcore/ --query
    /tmp/dumbselect.arq
    
    09:45:47 INFO  exec                 :: QUERY
      SELECT  ?subject ?predicate ?object
      WHERE
        { ?subject  ?predicate  ?object }
      LIMIT   25
    09:45:47 INFO  exec                 :: ALGEBRA
      (slice _ 25
        (project (?subject ?predicate ?object)
          (quadpattern (quad <urn:x-arq:DefaultGraphNode> ?subject
    ?predicate ?object))))
    09:45:47 INFO  exec                 :: TDB
      (slice _ 25
        (project (?subject ?predicate ?object)
          (quadpattern (quad <urn:x-arq:DefaultGraphNode> ?subject
    ?predicate ?object))))
    09:45:47 INFO  exec                 :: Execute ::   ?subject ?predicate 
?object
    --------------------------------
    | subject | predicate | object |
    ================================
    --------------------------------
    Time: 0.146 sec
    
    
    Ideally I would like to be able to test my usual queries in tdbquery
    (without adding some GRAPH information). I can't find a way to specify
    the UnionGraph pattern (no --graph argument like in tdbstats). Is
    there a way to make it work?
    
    Best,
    -- 
    Elie
    




Reply via email to