Re: R2RML/RML for CSV

On 8 October 2014 11:13, Anastasia Dimou <anastasia.dimou@ugent.be> wrote:
>
>>> Thanks! Progress so far is all in Github now. I've reorganized the
>>> files to match the "attempts/attempt-1" etc structure we've used for
>>> other use cases, so that [R2]RML can be contrasted with
>>> Mustache-family approaches, Direct Mapping(s) etc.
>>>
>>> https://github.com/w3c/csvw/tree/gh-pages/examples/tests/scenarios/events
>>>
>>> Here are the triples we get out so far...
>>>
>>>
>>> https://github.com/w3c/csvw/blob/gh-pages/examples/tests/scenarios/events/output/_output_attempt-1.nt.txt
>>>
>>> Here is the mapping file itself:
>>>
>>>
>>> https://github.com/w3c/csvw/blob/gh-pages/examples/tests/scenarios/events/attempts/attempt-1/mapping-events.rml.ttl
>>>
>>> There's an issue (Anastasia is investigating) in which the 'offers'
>>> and 'location' triples are not emitted, even though the Offer and
>>> Place entities and their properties are emitted.
>
> It was a matter of how the RML processor handled the mapping of referencing
> object maps when they are blank nodes.
> The issue is fixed (development branch) and now all triples are generated as
> expected.

Hey that's great - thanks for the quick fix. I've updated the WG
github repo with the actual output,

https://github.com/w3c/csvw/blob/gh-pages/examples/tests/scenarios/events/output/_output_attempt-1.nt.txt

... which matches the 10 triples per row expected output in

https://github.com/w3c/csvw/blob/gh-pages/examples/tests/scenarios/events/output/expected-triples.txt

It's tempting to express the goal triples in terms of SPARQL tests,
but I think for now it would be more useful to put time into trying
samples from other usecases, and expressing mappings in other
frameworks (e.g. mustache-like text based). Often we don't have super
explicit goal triples anyway.

cheers,

Dan

Received on Wednesday, 8 October 2014 11:10:20 UTC