@Bala Vignesh N V the issue is first() method returns a string not a Rdd. Subtract will works within two rdd's. So u should convert tagsheader to rdd by using parallelize. tags = sc.textFile("hdfs:///data/spark/genome-tags.csv") tagsheader = tags.first() header = sc.parallelize([tagsheader]) tagsdata = tags.subtract(header) ... View more