Spark has evolved considerably since the early days. Few new applications today use the Resilient Distributed Dataset (RDD), which have largely been replaced by DataFrames. In concert with the shift ...