Hardwood 1.0, being a light-weight, multi-threaded various, bypasses bloated legacy pipelines to course of a formidable 16.5 million rows per second throughout concurrent CPU cores.
On 22 June 2026, Hardwood launched as Common Availability (GA) model 1.0.0. Kickstarted by Java Champion Gunnar Morling, the creator of the One Billion Row Problem, the mission was constructed AI-natively over a five-month improvement cycle, although code evaluation and software program design remained completely human-owned.
Hardwood solves ‘heavyweight’ JVM constraints. Conventional Apache Parquet Java implementations, like parquet-java, introduce large dependency footprints, pulling in Hadoop and Avro, together with classpath vulnerabilities, whereas counting on a single-threaded reader. Written from scratch to discover totally different engineering trade-offs, Hardwood serves as a quick, less complicated, multi-threaded various for Java 21 or newer.
In contrast to older sequential implementations, Hardwood spreads Parquet web page decoding throughout all accessible CPU cores by default, saturating the host’s I/O and CPU bandwidth. The core framework includes a zero-mandatory-dependency profile to minimise supply-chain dangers; uncompressed or Gzip information require no third-party libraries. Non-compulsory codecs like LZ4, Snappy, Brotli, Zstd, or AWS S3 utilities connect as single-JAR extensions.
Filtered scans leverage branchless, batch-at-a-time analysis to minimise CPU department mispredictions. Information entry is break up into two clear layers: a structured Row Reader API for normal file mapping, and a batch-oriented Column Reader API engineered for high-throughput analytics pipelines.
In efficiency scans on 8 vCPUs, the reader achieved a throughput of 16.5 million rows per second. Model 1.0 additionally ships with an interactive text-based person interface (TUI) command-line software to examine file schemas and buildings immediately with out boilerplate code. Hardwood 1.0 is presently reader-only, with file-writing capabilities and Apache Arrow zero-copy interop on the roadmap.