I know I can use the df.compare to get the difference between the old and new prices like this: So The claim numbers and already paid amounts are 47000 lines of CPT codes and allowed/paid amounts.