W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2010

Re: project expression tests

From: Axel Polleres <axel.polleres@deri.org>
Date: Tue, 24 Aug 2010 14:46:36 +0100
Cc: "SPARQL Working Group" <public-rdf-dawg@w3.org>
Message-Id: <5F0FB64E-BADB-44C0-BD11-C22EB4E57D7D@deri.org>
To: Gregory Williams <greg@evilfunhouse.com>
On 12 Aug 2010, at 18:43, Gregory Williams wrote:

> I had a couple of comments about two of the project expression tests in CVS.
> 
> :projexp02:
> 
> I think the second expected query result should be using the plain literal "foobar" instead of "foobar"^^xsd:string, as the literal is untyped in projexp02.ttl.

Agree in principle, but see also below...

> 
> :projexp03:
> 
> projexp03.rq needs a '.' separating the two triple patterns.
> 

fixed that on CVS.

> Based on my understanding of XQuery F&O, I believe all the numerics in the query results should be of type xsd:integer. Right now the srx file has ?twice = "6"^^xsd:long, but the F&O document says "each operator is defined to operate on operands of the same type and return the same type" (with op:numeric-divide being the one exception).

That's an interesting point. How about SPARQL engines internally canonicalising datatpyes? is that an issue? What id the sum of two integers goes outside integer? Is that an issue? (probably not for these test cases, but in general?)


Another issue: There are no parentheses around project expressions in these test cases, I think we agreed on that, right?

Axel

> 
> 
> thanks,
> .greg
> 
> 
> 
Received on Tuesday, 24 August 2010 13:47:09 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:43 GMT