pyspark.streaming.DStream.countByValue#

DStream.countByValue()[source]#

Return a new DStream in which each RDD contains the counts of each distinct value in each RDD of this DStream.