Set this RDD’s storage level to persist its values across operations
after the first time it is computed. This can only be used to assign
a new storage level if the RDD does not have a storage level set yet.
If no storage level is specified defaults to (MEMORY_ONLY).
>>> rdd = sc.parallelize(["b", "a", "c"])