Hi Romain, all!
I looked into this, and have come up with a fix (commited on trunk as
rev 1240). I also added your code as a test (specifically for bound,
in "test/sparql_filter_bound.py").
However, this made me a little unconfortable. It was the second
problem I've encountered that stems from the change in Variable.py to
no longer keep its prefixing "?" internally. It seems the bindings
dict in "rdflib/sparql/sparqlOperators.py" for bound (and other
functions) contains Variable instances.
My fix was *not* to add e.g. an __eq__ method to Variable with logic
for this (allowing "?", and "$" too, possibly, since sparql allows
this - not n3 though?). I am not sure of the effects of doing that in
other places. (Also, I believe it is correct to keep the *name* and
not any sigil in Variable, so I'm not opposed to that change in
itself.)
Instead I changed, in "sparqlOperators.py", this function:
def __getQueryString(v) :
if isinstance(v,Variable) :
return v
elif queryString(v) :
return v
else :
return None
into:
def __getVariableName(v):
if isinstance(v, Variable):
return v
elif queryString(v):
return v[1:]
else:
return None
(and replaced all uses in the same file accordingly). This since
queryString already checks for the leading "?", so this was least
invasive and works.
I fear that this entire variable thing may be frail though, and that
it would be sound to investigate if we could normalize this all over
the code base. Good starting points: search for imports of
"rdflib.sparql._questChar", the use of "[1:]" and "?".
(On a related note, we should reasonably round up the different sparql
tests and normalize plus complete them. There is a bunch of these
"one-offs" right now. And in any event, many of these have regressed
since 2.4.0. As noted, changing
rdflib.sparql.Algebra.DAWG_DATASET_COMPLIANCE to False makes many of
them succeed, so that's a good place to start looking.)
Best regards,
Niklas
On 7/27/07, Romain Chantereau <[EMAIL PROTECTED]> wrote:
> Hi !
>
> It seems that there is a bug with the rdflib 2.4.0 sparql query, here an
> code example inspired from
> http://www.w3.org/TR/rdf-sparql-query/#func-bound :
>
> from rdflib import Literal, ConjunctiveGraph, Namespace, BNode, URIRef
>
> DC = Namespace(u"http://purl.org/dc/elements/1.1/";)
> FOAF = Namespace(u"http://xmlns.com/foaf/0.1/";)
>
> graph = ConjunctiveGraph()
> s = BNode()
> graph.add((s, FOAF['givenName'], Literal('Alice')))
> b = BNode()
> graph.add((b, FOAF['givenName'], Literal('Bob')))
> graph.add((b, DC['date'], Literal("2005-04-04T04:04:04Z")))
>
> print graph.query("""PREFIX foaf: <http://xmlns.com/foaf/0.1/>
> PREFIX dc: <http://purl.org/dc/elements/1.1/>
> PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
> SELECT ?name
> WHERE { ?x foaf:givenName ?name .
> OPTIONAL { ?x dc:date ?date } .
> FILTER ( bound(?date) ) }""").serialize('python')
>
> I think the expected result is [Literal('Bob')], but the result I have
> is [].
>
> Thanks for your really great job,
> --
> Romain Chantereau <[EMAIL PROTECTED]>
>
> _______________________________________________
> Dev mailing list
> Dev@rdflib.net
> http://rdflib.net/mailman/listinfo/dev
>
_______________________________________________
Dev mailing list
Dev@rdflib.net
http://rdflib.net/mailman/listinfo/dev
