Skip to content

hyparam/hyparquet-compressors

Repository files navigation

hyparquet decompressors

hyparquet parakeets

npm workflow status mit license coverage

This package exports a compressors object intended to be passed into hyparquet.

Apache Parquet is a popular columnar storage format that is widely used in data engineering, data science, and machine learning applications for efficiently storing and processing large datasets. It supports a number of different compression formats, but most parquet files use snappy compression.

The hyparquet library by default only supports uncompressed and snappy compressed files. The hyparquet-compressors package extends support for all legal parquet compression formats.

The hyparquet-compressors package works in both node.js and the browser. Uses js and wasm packages, no system dependencies.

Usage

import { parquetRead } from 'hyparquet'
import { compressors } from 'hyparquet-compressors'

await parquetRead({ file, compressors, onComplete: console.log })

See hyparquet repo for further info.

Compression formats

Parquet compression types supported with hyparquet-compressors:

  • Uncompressed
  • Snappy
  • Gzip
  • LZO
  • Brotli
  • LZ4
  • ZSTD
  • LZ4_RAW

Snappy

Snappy compression uses hysnappy for fast snappy decompression using minimal wasm.

Gzip

New gzip implementation adapted from fflate. Includes modifications to handle repeated back-to-back gzip streams that sometimes occur in parquet files (but was not supported by fflate).

Brotli

Includes a minimal port of brotli.js which pre-compresses the brotli dictionary using gzip to minimize the distribution bundle size.

LZ4

New LZ4 implementation includes support for legacy hadoop LZ4 frame format used on some old parquet files.

Zstd

Uses fzstd for Zstandard decompression.

Bundle size

File Size
hyparquet-compressors.min.js 116.1kb
hyparquet-compressors.min.js.gz 75.2kb

References