Re: Test cases (batch 2) from Peter

On 18/04/2017 23:55, Peter F. Patel-Schneider wrote:
> I stupidly missed the "for the datatypes supported by SPARQL 1.1, is not an
> ill-typed literal" so there should probably be three violations.  I've
> updated the test.

Ok thanks.

>
> However, there is no pointer to just what are the supported SPARQL 1.1
> datatypes, and the situation with respect to supported datatypes in SPARQL
> 1.1 is somewhat murky.  In Section 17.1 of SPARQL 1.1 Query Language there
> is a list of datatypes that provide typing for SPARQL operators.  However,
> there is no mention of "supported" or "support" in this section.  In Section
> 3.3 there is "In addition to numeric types, SPARQL supports types
> xsd:string, xsd:boolean and xsd:dateTime".  There are also a few examples of
> supported datatypes.  This makes it difficult to determine with precision
> what the datatypes supported by SPARQL 1.1 are.
>
> It thus would probably be a good idea to provide a better description of
> what datatypes are meant here, something along the lines of "the XML Schema
> datatypes listed in Section 17.1 of SPARQL 1.1 Query Language".

Section 17.1 does not appear to be complete, e.g. xsd:date is missing 
yet it should be tested by sh:datatype. The absence of a crisp place in 
the SPARQL spec probably led us to stay rather vague on this. I have put 
the question on whether this requires changes on the agenda for the WG 
meeting today.

Holger


>
>
>
> On 04/18/2017 03:02 AM, Holger Knublauch wrote:
>> Hi Peter,
>>
>> thanks for another batch of tests, as submitted here:
>>
>> https://github.com/w3c/data-shapes/pull/52
>>
>> I again had to fix a syntax error - please check that the files are at least
>> valid Turtle before submitting. Thanks.
>>
>> The test /core/property/datatype-ill-formed does not look correct to me. You
>> expect that only the xsd:integer 55 is invalid, but IMHO "300"^^xsd:byte and
>> "c"^^xsd:byte are also invalid literals for xsd:byte.
>> https://www.w3.org/TR/xmlschema-2/#byte sets its range between -128 and 127,
>> so 300 would be invalid. Likewise "c" would be invalid.
>>
>> Please clarify.
>>
>> Holger
>>

Received on Wednesday, 19 April 2017 00:24:47 UTC