Difference in caching RDD vs. caching DataFrame in Spark

Question

In the book "Spark - The Definitive Guide" it is written:

When we cache an RDD, we cache the actual, physical data (i.e, the bits). The bits. W