FastLanes

Like Parquet, but with 40% better compression and 40× faster decoding.

FastLanes features:

• Fully SIMD-friendly with zero explicit SIMD instructions

• Zero dependencies:

  • no external libraries required

Documentation

• Typing System
• File Format Specification (PDF)

Getting Started

Python

import pyfastlanes

# Connect to FastLanes
conn = pyfastlanes.connect()

# Convert a CSV directory to FastLanes format
conn.inline_footer().read_csv("path/to/csv_dir").to_fls("data.fls")

# Read back and write to CSV
reader = conn.read_fls("data.fls")
reader.to_csv("decoded.csv")

C++

Add FastLanes as a dependency via CMake:

include(FetchContent)
FetchContent_Declare(
        fastlanes
        GIT_REPOSITORY https://github.com/cwida/FastLanes.git
        GIT_TAG dev
)
FetchContent_MakeAvailable(fastlanes)

add_executable(example example.cpp)
target_link_libraries(example PRIVATE FastLanes)

Example usage:

#include "fastlanes.hpp"

int main() {
    fastlanes::Connection conn;
    conn.read_csv("data/csv_dir").to_fls("data.fls");

    auto reader = fastlanes::Connection().read_fls("data.fls");
    reader->to_csv("decoded.csv");
    return EXIT_SUCCESS;
}

Rust

Add FastLanes Rust bindings to your Cargo.toml:

[dependencies]
fls-rs = { path = "./rust" }

use anyhow::Result;
use fls_rs::connect;

fn main() -> Result {
    let mut conn = connect();
    conn.inline_footer()
        .read_csv("data/csv_dir")
        .to_fls("data.fls");

    conn.read_fls("data.fls")
        .to_csv("decoded.csv");
    Ok(())
}

Coming Soon

• CUDA support for FastLanes CUDA reader

Publications

• Azim Afroozeh & Peter Boncz, “The FastLanes Compression Layout: Decoding > 100 Billion Integers per Second with Scalar Code,” PVLDB, 16(9): 2132–2144, May 2023

  • Read the paper
  • Source code

• Azim Afroozeh, Lotte Felius & Peter Boncz, “Accelerating GPU Data Processing Using FastLanes Compression,” DaMoN ’24, Proceedings of the 20th International Workshop on Data Management on New Hardware, Santiago, Chile, June 2024

  • Read the paper
  • Source code

• Azim Afroozeh, Leonardo Kuffó & Peter Boncz, “ALP: Adaptive Lossless Floating-Point Compression,” SIGMOD ’24, ACM SIGMOD, June 2024

  • Read the paper
  • Source code

• Sven Hielke Hepkema, Azim Afroozeh, Charlotte Felius, Peter Boncz & Stefan Manegold, “G‑ALP: Rethinking Light‑weight Encodings for GPUs,” DaMoN ’25, July 2025

  • Read the paper
  • Source code

How to Cite

If you use FastLanes in your research or projects, please cite:

…

License

This project is released under the MIT License.

Contributing

We welcome contributions to FastLanes!
Please see CONTRIBUTING.md for guidelines on how to get started.

Join Our Community

Come discuss FastLanes, share feedback, and help shape the future of data formats on Discord: