Re: [BioRDF] RDF query languages

[Seaborne, Andy]

Xiaoshu Wang wrote:
Of course, the best bet would likely be to go with 
SPARQL[1], the RDF 
Query Language defined by the W3C RDF Data Access Working Group 
(DAWG)[2].
The DAWG produces specifications for both the query language and a 
protocol to issue queries and receive results (either as an 
RDF graph 
or as an XML representation of a result set).
+1

http://esw.w3.org/topic/DawgShows additionally might be of 
use to some interested in learning more about SPARQL.

+1.

I didn't know the XQuery in depth.  But my intuiation is that it must be
designed to query against the semantics of XML, which by specification only
specifies the semantics of document structure. But the same set of
information can be represented in multiple ways with RDF/XML.  Thus, any
attempt to derive the RDF semantics by querying the structure of a document
runs the risk of miss-"interpretation" or misinterpretation because the
required another level of indirection is not standardized.  GRDDL can work
because it is the same person who will authorize the HTML page and XSLT
transformation sheet.  The same cannot be said with any RDF document.
Design an RDF-aware application based on XQuery will not be portable to
other knowledge platform.  In addition, what if the RDF model is encoded in
language of non-RDF/XML, for instance the Notation-3 family?

My point is, as always, use the right tool to do the right thing.  Don't
heck unless there is no other way around.  

Xiaoshu

I translated the queries from

  http://www.isb-sib.ch/~ejain/expasy4j-webng/query.html#examples

into SPARQL:

  http://esw.w3.org/topic/LifeSciencesQueries/SPARQL


Notes:

- Example 4 uses COUNT which isn’t in SPARQL.
- It appears that ex9 has no answers on the 10e6 triple 7.0a UniProt extract
- Similarly, none of the queries have any results on the 1e6 triple 7.0a 
extract if I have prepared the data correctly.

        Andy