zpq

About

A small Unix command-line tool for Parquet that queries, filters, aggregates and rewrites files on local disk or in S3, and prints NDJSON, CSV or Parquet for the next tool in a pipe. It reads footers and statistics first and touches only the pages a query needs, keeps DECIMAL values as exact integers, and returns rows in the same order at any thread count. It is written in Zig.