I’m currently struggling with spark checkpoints and trying to understand what’s the difference between DataFrame and RDD checkpoints.

To make it clear: after DataFrame.checkpoint() spark creates file on hdfs, and it is something like hdfs:///sdfasd-dfasdf-dfasdf/rdd-2, – so it saves RDD of DataFrame. The same thing saves after checkpointing RDD.
So, the main question: does this files differ? and will spark use this checkpoint for dataframe if we made RDD.checkpoint() (if they are the same files).

And also I have some side question: for RDD there are getCheckpointFile() method which allows to get checkpoint file full path. Are there any chance to get the same for DataFrame? Because there are no box method.