Simon, let me go back to my original question:

This is a simplified version of the BIBFRAME data:

:instance27 bf:publication _:genid4 .
_:genid4 bf:providerName _:genid5 .
_:genid5 bf:label "Morrow" .

It perhaps makes more sense to look at it in the RDF/XML, before converting it to triples:

 <bf:publication>
      <bf:Provider>
        <bf:providerName>
          <bf:Organization>
            <bf:label>Morrow</bf:label>
          </bf:Organization>
        </bf:providerName>
      </bf:Provider>
    </bf:publication>

Assuming this is behind a SPARQL end-point (or an imitation thereof), to retrieve the bf:label for the bf:providerName, you need to be able to make a sparql query that finds the subject of a triple with predicate bf:providerName, then uses that IRI as the subject of the query seeking the bf:label.  It appears, however, that this does not work when the object later used as a subject is a bnode.

The query below, which I thought would work, fails to return anything:

PREFIX bf: <http://bibframe.org/vocab/>
SELECT ?subject ?label
    WHERE { ?subject bf:providerName ?bnode .
                 ?bnode bf:label ?label . }

EXCEPT when the blank nodes are converted to actual IRIs (which Virtuoso does when it interprets RDF/XML). That is, with actual bnodes, the query fails; with skolemized [1] identifiers, it succeeds.

I think there is a hint in this answer online: " with a SPARQL-compliant system it is not possible to refer to the identifier of a blank node in the queried graph from within a SPARQL query pattern. " [2] However, I can't find an example that looks like the BIBFRAME case, and that particular "explanation" isn't crystal clear to me.

(Maybe I should post a query to the "answers" forum.)

kc
[1] https://en.wikipedia.org/wiki/Skolem_normal_form
[2] http://answers.semanticweb.com/questions/1673/skolem-syntax-for-blank-nodes-in-sparql



On 7/15/14, 5:39 PM, Simon Spero wrote:
[log in to unmask]" type="cite">
It is generally better to avoid blank nodes in SPARQL queries, since it is usually  clearer if you use  ?y instead of _:y . 

Simon


On Tue, Jul 15, 2014 at 8:37 PM, Simon Spero <[log in to unmask]> wrote:
That's would be the same query (and same basic graph pattern). 


On Tue, Jul 15, 2014 at 7:18 PM, Karen Coyle <[log in to unmask]> wrote:
On 7/15/14, 10:47 AM, Simon Spero wrote:
4. There is no problem with a SPARQL  query binding a variable to a blank node; however it can be difficult to use the value in second query.
Simon, I admit that I struggle often with the terseness of your prose. Is this #4 your statement of the problem that I ran into, as per my original post? In other words, is

Where {
X p _y .
_y p Z .
}

using the value in a second query? If not, does one of your other points address this?

thx,

kc

--
Karen Coyle
[log in to unmask] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet



-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet