ALP: A New Standard for Adaptive Lossless Floating-Point Compression
ALP: Adaptive lossless floating-point compression
We present ALP, a state-of-the-art lossless compression algorithm designed specifically for IEEE 754 floating-point data. By exploiting patterns in decimal and high-precision numbers, ALP achieves superior performance in decoding speed, compression ratio, and encoding speed. Our approach integrates seamlessly with systems like DuckDB and FastLanes, earning the SIGMOD Best Artifact Award for its reproducibility and effectiveness.
The right part is left uncompressed, as it is required to preserve high precision and is often highly random and incompressible.