Re: SOURCE in N3QL

On Sep 27, 2004, at 4:07 PM, Seaborne, Andy wrote:

> PS Maybe we should extend the test case vocabulary to allow for
> different query languages, as well as the request for data syntax
> annotations.  Such a syntax property that can be applied to a 
> individual
> query or to a whole manifest.  Then there can be a record all your 
> Euler
> tests for the WG.

+1

I think we already hacked something like that in the old RDF-query 
testing framework in the past

http://www.w3.org/2003/03/rdfqr-tests/summary.html#manifest

running at

http://www.w3.org/2003/03/rdfqr-tests/test/

worth to diff the old and new manifest format and get the best of of it?

Yours

Alberto

>
> -------- Original Message --------
>> From: Jos De_Roo <mailto:jos.deroo@agfa.com>
>> Date: 25 September 2004 19:56
>>
>> Andy,
>> Eric,
>>
>> I made 2 test cases
>>
>> 1/
>> data:  http://eulersharp.sourceforge.net/2004/04test/D1.n3
>> query: http://eulersharp.sourceforge.net/2004/04test/D1Q.n3
>> gives: http://eulersharp.sourceforge.net/2004/04test/D1E.n3
>>
>> which is an empty answer
>> because the bnodes in D2.n3 could be some different things
>>
>>
>> 2/
>> data:  http://eulersharp.sourceforge.net/2004/04test/D2.n3
>> query: http://eulersharp.sourceforge.net/2004/04test/D2Q.n3
>> gives: http://eulersharp.sourceforge.net/2004/04test/D1E.n3
>>
>> which is 2 answers
>> because the bnodes came from the same graph D2.n3
>> and so they unify
>>
>> at least that is my current understanding about relabeling
>> bnodes and unification :)
>>
>>
>> jos
>>
>> --
>> Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
>>
>>
>>
>>
>> "Seaborne, Andy" <andy.seaborne@hp.com>
>> 25/09/2004 17:01
>>
>>
>>         To:     Jos De_Roo/AMDUS/MOR/Agfa-NV/BE/BAYER@AGFA, "Eric
>> Prud'hommeaux" <eric@w3.org>
>>         cc:
>>         Subject:        SOURCE in N3QL
>>
>>
>> Jos,
>>
>> I tried to explain this to Eric in an email and I have a question:
>>
>>> well, I thought that
>>>
>>>   SELECT ?x ?y ?n
>>>   WHERE
>>>     (?x foaf:knows ?y)
>>>     SOURCE <D2.n3> (?x foaf:age ?n)
>>>     ...
>>>
>>> nicely maps to
>>>
>>>   []
>>>   q:select {(?x ?y ?n) a q:Answer};
>>>   q:where  {?x foaf:knows ?y.
>>>             <D2.n3>.log:semantics log:includes {?x foaf:age ?n}}.
>>>             ...
>>
>> ?x is a bNode.
>>
>> My thought:
>>
>> ----
>> The query assumes that <D2.n3> has the same bNode as the outer context
>> "?x
>> foaf:knows ?y" so there may be there is an assumption that the global
>> graph is the RDF merge of the named components.
>> ----
>>
>> Is my thought right or wrong or misguided?
>>
>> Which ever, this is beginning to give a possible theoretical
> foundation
>> for SOURCE.
>>
>> Andy
>

Received on Monday, 27 September 2004 16:56:36 UTC