How to Work with Mixed Data Types in Parquet Files Using PyArrow and Pandas for Efficient Data Storage