Unity Catalog
| Unity Catalog | |
|---|---|
| Developer | Unity Catalog community (LF AI & Data Foundation) |
| Release | June 12, 2024 (open source) |
| Stable release | 0.5.1
/ July 18, 2026 |
| Written in | Java, Python, TypeScript, Scala |
| Operating system | Cross-platform |
| Type | Metadata catalog, data governance |
| License | Apache License 2.0 |
| Website | www |
| Repository | github |
Unity Catalog is a metadata and governance catalog for data and artificial intelligence assets, released under the Apache License. It includes a unified namespace for tables, unstructured data volumes, and functions, with APIs that can be accessed with multiple compute engines. It also manages table formats including Delta Lake and Apache Iceberg.[1][2] The source code for the project is part of the LF AI & Data Foundation, an open source nonprofit software foundation.[3][4] The catalog was originally developed at Databricks, which also offers it as a managed service.[5]
Unity Catalog works by organizing data and AI assets in a hierarchical namespace consisting of catalogs, schemas, and other securable objects.[2][6][7] It manages structured tables, supporting formats such as Delta Lake, Apache Iceberg, Apache Parquet, CSV, and JSON; and unstructured data through volumes; and machine-learning assets such as functions and models.[2][8]
The governance features in the project include credential vending, a feature where the catalog server issues scoped credentials to clients accessing underlying cloud storage, and support for the Apache Iceberg REST Catalog API and Apache Hive metastore API.[2][8] By acting as a layer that exposes open APIs, implementation provides an interface for external clients, such as engines like Apache Spark to read tables, volumes, and functions that are managed by the catalog.[9]
Unity Catalog leverages an OpenAPI-based REST specification and an reference server, with client libraries and SDKs.[8][6] The server is compatible with the popular Apache Iceberg REST Catalog and provides compatibility with the Apache Hive metastore API.[2][8]
Informasi ini disarikan dari Wikipedia dan disajikan kembali untuk tujuan edukasi. Konten tersedia di bawah lisensi CC BY-SA 3.0. Kami tidak bertanggung jawab atas ketidakakuratan data yang bersumber dari kontribusi publik tersebut.