I have a libsvm file that I want to load into Spark and then encrypt it. I want to iterate over every element in the features to apply my encrypt function, but there doesn’t seem to be any way to manipulate a VectorUDT which is the class of the features after using spark.read.format("libsvm").load(..)