how to read million records in jupyter notebook