HeadlinesBriefing favicon HeadlinesBriefing.com

Open‑Source F3 File Format Tackles Parquet Limits

Hacker News •
×

The research team behind the SIGMOD 2026 paper released F3, an open‑source columnar file format built to address layout inefficiencies in legacy formats such as Parquet. By embedding Wasm decoders directly in each file, F3 promises cross‑platform compatibility without relying on external libraries. The codebase lives on GitHub under the future‑file‑format organization and is licensed MIT for early adopters.

The repository includes a flat‑buffer schema, a proof‑of‑concept package, and a suite of benchmarks that replicate the experiments reported in the paper. Build scripts target Debian 12 on Intel hardware, and the fff‑bench directory holds micro and end‑to‑end tests comparing storage layout and decoding speed against ORC and Parquet, with detailed performance logs. Results show measurable gains in both size and runtime.

While the authors caution that F3 remains a prototype and is not recommended for production workloads, the project demonstrates a viable path toward self‑describing, extensible storage. Embedding decoders adds only a few kilobytes per file, eliminating version mismatches when native libraries lag behind. The code now serves as a reference implementation for future file‑format research.