I\'m trying to implement a NLP-Pipeline using Spark/Scala.
Right now I face the difficulty of subtracting one collection (implemented as Dataframes) from another - bo