RDFUnitStaticWrapper wrong validation

22 views
Skip to first unread message

Guido Zuccarelli

unread,
Nov 17, 2014, 10:08:21 AM11/17/14
to rdf...@googlegroups.com
I downloaded the Java library in order to test my dataset.
Then I tried to validate it by executing the following code: http://pastebin.com/tyhX2w2F which uses the schema ontology.
The execution is too fast, 10 seconds for a dataset of 3 GB.
This is the result:
http://pastebin.com/5YsMSzAL
0 errors. I manually found in the first check many range errors, so I think it is wrong.
Am I using the library incorrectly?

Dimitris Kontokostas

unread,
Nov 17, 2014, 10:26:45 AM11/17/14
to Guido Zuccarelli, rdf...@googlegroups.com
Hi Guido,

If this is the actual code, I do not see where you load the data. 
It looks like RDFUnit validates an empty jena model and that is why it is so fast fast with no errors.

Is this the case?

BTW, you'd better use the following URI for schema.org
This is also used from LOV
and it should be more recent than the one you use (but not 100% sure)


Best,
Dimitris

--
You received this message because you are subscribed to the Google Groups "RDFUnit" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rdfunit+u...@googlegroups.com.
To post to this group, send email to rdf...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rdfunit/c627e1e2-e273-4403-acca-c81c4c07b4fb%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Kontokostas Dimitris

Dimitris Kontokostas

unread,
Nov 17, 2014, 10:29:17 AM11/17/14
to Guido Zuccarelli, rdfunit
[update]

Just saw that you load data in imputModel but are you sure that the data is loaded before RDFUnit runs
Such a fast behavior usually comes from an empty model 
--
Kontokostas Dimitris

Guido Zuccarelli

unread,
Nov 17, 2014, 10:59:59 AM11/17/14
to rdf...@googlegroups.com, guidozu...@gmail.com
You were right, thank you!
This is the sentence that I should use dataset.getNamedModel("urn:x-arq:UnionGraph")
Reply all
Reply to author
Forward
0 new messages