KD2R aims to discover composite keys in one or several RDF
data source. These data conform to the same OWL ontology and are described in RDF files for which the UNA is fulfilled. KD2R takes into account the properties that the RDF files may have: incompleteness and multi-valuation.
Since the data may be numerous, KD2R discovers maximal undetermined/non keys (exploring only a part of the dataset) that are used to compute keys. Keys that are discovered using different datasets are merged.
