Apache Parquet

github.com ↗

- On-disk columnar representation of data compatible with Pandas, Hadoop-based systems, etc..