Fastest way to iterate over 70 million rows in pandas dataframe

Question

I have two pandas dataframes bookmarks and ratings where columns are respectively :

• id_profile, id_item,