On 22/09/11 21:15, David Jordan wrote:
Relative to the second topic in this this posting, I got my query working.
The following query works:
select distinct ?subjid ?diag
FROM<http://purl.bioontology.org/ontology/HOM_ICD9_inferred/>
FROM<http://www.sas.com/hls/hoa/patient/>
where {
?subj p:hasDiagnosis ?diag .
?subj p:hasPatientID ?subjid .
FILTER( EXISTS { ?diag rdf:type HOM_ICD9:HOM_ICD_1224 } )
}
However, I thought using
--set tdb:unionDefaultGraph=true
would put the two graphs into the default graph (those are my only 2 graphs
right now),
not requiring me to have the FROM clause at all. But with that, I had no
results.
So this seems like an inconsistency to me.
David,
You don't need FROM or FROM NAMED at all. They exist to describe the
dataset and if there isn't a description (the most common case) the
dataset is just what is in the TDB datastore.
So
--set tdb:unionDefaultGraph=true
should work.
With TDB, use FROM/FROM NAMED only if you want to pick a subset of the
graph out of a large collection of graphs.
Is there some way we can share a test case here? i.e a small set of
data and queries so I can reproduce your setup in my development system?
Andy
