python pandas merge - finding difference between two data frames (NaN issue)

Question

I have a customer profile csv (containing 450,000 records/customers and 57 attributes) which arrives every week. I\'d like to filter the new dataframe for rows that are new