How to split out _corrupt_record from Spark DataFrame in Scala?

Question

In Spark with Scala I need to separate out incoming data that have a _corrupt_record.

I have the following code:

Here I read in the data into a DF