Removing Duplicates from Pandas DataFrame with Keep First Event Only on fast_order Category While Removing Duplicates from All Other Categories