Hello
We've done a lot of experiments with jena-solr today. We loaded a OntModel
based on OWL_MICRO, with data in memory, loading a bunch of N3 and OWL
files.
First, we've loaded a N3 file into a TDB and indexed it, according to the
documentation. Then we removed the usage of TDB and instead used that model
with the SOLR index. It still could solve SPARQL with text:query correcly.
(1) We tried then to create a query mixing text:query and other assertions
(loaded on n3). Example:
SELECT * {
?a a MyClass.
?a text:query "my text".
}
This case above retrieves all MyClass instances a certain number of times,
which happens to be the number of instances that solr retrieves for
querying "my text". Also, the SOLR log shows that there has been this exact
number of queries being requested.
If I switch places, it will work as intended:
SELECT * {
?a text:query "my text".
?a a MyClass.
}
(2) The other experiment we did was to add a inference axiom that used the
solr index:
text:query someValuesFrom "my text" SubClassOf MyFancyClass
which is well within the capabilities of OWL_MICRO. However, the query
SELECT * {
?a a MyFancyClass.
}
yielded no results. The Solr log showed no requisition either.
Now the questions =)
I can fix (1) by using FILTER, but I guess the correct behaviour in this
case would be that the query executor add something to the solr query
specifying the results from MyClass, is that so?
I couldn't figure why (2) didn't work.
Thanks!
--
diogo patrão
