How do Spark's cache and persist operations work? Would you classify them as actions or transformations?

cache() is a shorthand for persist(StorageLevel.MEMORY_ONLY). Both methods are lazy transformations that tag an RDD/DataFrame to be stored in the specified storage level. They return the same logical plan but with a persistence flag. The actual materialization happens when an action (e.g., count, collect) forces the RDD to be computed. Therefore, cache() and persist() are considered transformations, not actions.