Вы можете использовать flatMap следующим образом:
rdd = sc.parallelize([(2, [199.99, 250.0, 129.99]),
(4, [49.98, 299.95, 150.0, 199.92]),
(8, [179.97, 299.95, 199.92, 50.0]),
(10, [199.99, 99.96, 129.99, 21.99, 199.99]),
(12, [299.98, 100.0, 149.94, 499.95, 250.0])])
print rdd.flatMap(lambda x: [(x[0], y) for y in x[1]])\
.sortBy(lambda x: (x[0], x[1])).collect()
[(2, 129,99), (2, 199,99), (2, 250,0), (4, 49,98), (4, 150,0), (4,
199,92), (4, 299,95), (8, 50,0), (8, 179,97), (8, 199,92), (8, 299,95), (10, 21,99), (10, 99,96), (10, 129,99) , (10, 199,99), (10, 199,99), (12, 100,0), (12, 149,94), (12, 250,0), (12, 299,98), (12, 499,95)]