How can I cast/extract nested column field from scala DataFrame to a case class most efficiently?

Question

I have a case class like :

case class PA( id: Long, name: String, level: Int, ......., path: Seq[ParentPath] )

and ParentPath class is like :