Closed KonradHoeffner closed 3 years ago
SELECT ?x (COUNT(?userx) AS ?usagex) ?y (COUNT(?usery) AS ?usagey) (COUNT(?userx)-COUNT(?usery) AS ?diff)
{
?x a meta:Feature; rdfs:label ?labelx.
?y a meta:Feature; rdfs:label ?labely.
FILTER(?x!=?y AND ?x<?y).
FILTER(STR(LCASE(?labelx))=STR(LCASE(?labely)))
?userx ?p ?x.
?usery ?q ?y.
} GROUP BY ?x ?y
As the usage numbers are exactly the same, I could just remove the ones without the "Feature" prefix.
However those without the prefix have the English label tag, contain a HITO close match to a nonexistant resource and are defined by the BB ontology.
Those with the prefix are a typical feature of another class and are annotated by an axiom with tripel page.
So it needs a real merge.
Many features exist twice in the BB, e.g.
bb:ProvideCatalogsAndOtherMeansForCodingOfNursingDiagnosis vs bb:FeatureProvideCatalogsAndOtherMeansForCodingOfNursingDiagnosis
Find out which of those are used and merge them.