Re: rdb2rdf test cases

Hi Richard

On 21/07/2011 18:08, Richard Cyganiak wrote:
> Nice.
> Is there some way of creating a test case that's supposed to fail? For example, an invalid mapping file where some required property is missing?
Yes, there is a way of creating a "incorrect" test case ... I'll create 
them  ;)
You already suggested some of them here [1], I'll create tcs for the 
but, what would be the output of these "incorrect" test cases, an empty 
result with (or without) a error message?
One option is an empty result and an error message pointing out what 
R2RML section is wrong/missing.
what do you think?


> Richard
> On 21 Jul 2011, at 16:41, Boris Villazon-Terrazas wrote:
>> Hi all
>> I was modifying (including more) test cases we have in the cvs repository [0].
>> Now they include a manifest file that describes the elements of the TCs.
>> The manifest file follows the vocabulary we created in deri vocab [1]. You can also find the complete vocabulary here [2].
>> I renamed the result files, mapped*.trig, for mapped*.nq following Richard's suggestions.
>> I also renamed the directGraph files *.ttl to *nt .
>> Thanks to Nuno, I had to update less R2RML files to the current specification.
>> However, I'll need to update the "specificationReference" property of the descriptions.
>> Boris
>> [0]
>> [1]
>> [2]

Received on Thursday, 21 July 2011 20:50:41 UTC