Gunnar Morling • 4/2/2026

Hardwood Reaches Beta: S3, Predicate Push-Down, CLI, and More

This article announces the beta release of Hardwood, a minimal-dependency Apache Parquet parser. Key features include an S3 backend for direct parsing from object storage (S3, R2, GCS) without downloading, predicate push-down to reduce network IO, Avro bindings, and a CLI for inspecting Parquet files. The project emphasizes a small footprint, using Java's built-in HTTP client and custom SigV4 signing instead of heavy SDKs. Authentication supports simple keys or full AWS credential chains via an optional module. A new website with documentation is also launched.

0 comments

#cli #S3 #Parquet