i've not worked on or seen datastage, but would it be possible to reverse engineer the transformations by comparing the source & result? pyspark does not have any such built-in methods afaik.