add support for blank nodes
cweiske opened this issue · comments
Blank nodes are treated as strings currently, leading to duplicate blank node names when importing multiple RDF files (and thus data confusion).
The problem lies in Erfurt_Store_Adapter_Virtuoso::addStatement() which adds triples one by one, instead of storing them together so that real blank nodes can be used.
Not sure this can all be fixed completely, but I noticed I could solve a SPARQL query matching bnodes in 1.1 (see http://lists.w3.org/Archives/Public/public-webid/2012Jul/0055.html for a description of the problem) by applying the following change : olberger@4142a14
Maybe that just addresses the inserting into Virtuoso and not the caching stuff...
Hope this helps.
feature branch bnodes is maybe interesting for you:
https://github.com/AKSW/Erfurt/tree/feature/bnodes