ComparableDataset.scala

val comparableDataset = kvKDataset.as("l")
      .joinWith(
        kvKDataset.as("r"),
        $"l.adresV" === $"r.adresV" && $"l.postcodePlaatsV" === $"r.postcodePlaatsV" && $"l.dossierNummer" =!= $"r.dossierNummer"
      ).map {
        case (left, right) => (left, right, Vectors.dense(left.distance(right).toArray))
      }
      .toDF("left", "right", "features")
      .as[(KvKRecord, KvKRecord, Vector)]