- From: Holger Knublauch <holger@topquadrant.com>
- Date: Fri, 8 Jul 2016 18:43:00 +1000
- To: public-data-shapes-wg@w3.org
I had looked into this topic a while ago. My proposal is to keep the validation of ill-formed literals and sh:datatype separate, i.e. sh:datatype should not check for well-formedness. The reasons for that are two-fold: a) doing this extra test would have a performance penalty for such a frequently used operation. The test is not trivial, e.g. for xsd:dateTime. b) I see no way in standard SPARQL to express such a test. The best thing we came up with is to use a higher-level function to invoke the type casts such as xsd:integer("a"), following by a bound(?result) check. But such a higher-level "invoke" function does not exist in SPARQL 1.1 My preference is to leave this capability to extensions, i.e. close the ticket without action. Holger On 29/04/2016 19:59, RDF Data Shapes Working Group Issue Tracker wrote: > shapes-ISSUE-158 (ill-typed literals and sh:datatype): ill-typed literals do not always trigger a validation result [SHACL - Core] > > http://www.w3.org/2014/data-shapes/track/issues/158 > > Raised by: Peter Patel-Schneider > On product: SHACL - Core > > The node ex:i1 in > ex:i1 ex:p "c"^^xsd:integer > validates against > ex:s1 a sh:Shape ; > sh:property [ a sh:PropertyConstraint ; sh:predicate ex:p ; sh:datatype xsd:integer ] . > > I think that it should not validate because "c" is not in the lexical space of the datatype xsd:integer. > > >
Received on Friday, 8 July 2016 08:43:35 UTC