RDD = sc.parallelize([1, 2, 3, 4, 6, 5, 4, 3])
print RDD.distinct().collect()