Unexpected literal values

[mpagni12]

I found some "NaN"^^xsd:float values by running:

SELECT DISTINCT ?s
WHERE { 
    ?s ?p ?o .
    FILTER( STR( ?o ) = "NaN" )
} 
LIMIT 10

In fact, this is perfectly valid RDF (https://www.w3.org/TR/xmlschema11-2/). I have learned something new here, thank you. However I doubt that this added complication is terribly necessary ...

[oolonek]

To keep but replace by ND or Unknown as a string.

[mpagni12]

A mixture of strings and numbers is not a good ide, when one is post-processing of a lot of data using R, for example. I would simply suppress the triple that contain the NaN