Transactional catalog for data lakes with Git-like branching, commits, and Iceberg table support
License
Apache-2.0
Languages
- Java
- Scala
- Kotlin

About Project Nessie
Project Nessie is a transactional catalog for data lakes with Git-like semantics. Instead of a single mutable catalog, it versions table metadata as commits, so data teams can branch, tag, and roll back the entire state of their tables and views as one atomic unit.
Nessie manages Iceberg tables and views and connects to a broad set of engines, including Spark, Flink, Hive, Presto, and Trino through Iceberg. Bearer-token authentication backed by OpenID Connect can be switched on to control access.
The catalog runs as a lightweight server, and a CLI plus client libraries drive branching and commit workflows from data pipelines.
Key features
- Versioned table metadata as Git-like commits
- Branch, tag, and roll back tables and views atomically
- Manages Iceberg tables and views
- Connects to Spark, Flink, Hive, Presto, and Trino via Iceberg
- OpenID Connect bearer-token authentication
Details
- First released
- 2020
- Type
- Transactional data lake catalog
- Language
- Java
- Storage
- Versioned table metadata
- Compatibility
- Iceberg · Spark · Flink · Trino
- Deployment
- Self-hosted · Docker
