I am dealing with terabytes of mildly noisy data broken up into ~20 GB dataframes. Maybe 20 in each 2 million rows has bad data in certain columns. I want a quick forced type ca
