Unity Catalog is Databricks' unified governance layer for data and AI assets, built on a metastore and a three-level namespace (catalog → schema → object). It standardizes access control, auditing, lineage, and sharing across every workspace and compute type, and as of 2026 extends the same governance model to AI assets such as models, functions, and connections. A useful mental model is "secure by default": querying an object typically requires both container privileges (USE CATALOG + USE SCHEMA) and an object privilege (like SELECT), while governance objects (storage credentials, external locations) gate all cloud storage access. Since September 30 2025, Unity Catalog is mandatory for all new workspaces.
What This Cheat Sheet Covers
This topic spans 18 focused tables and 172 indexed concepts. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.
Table 1: Core Namespace Objects
| Object | Example | Description |
|---|---|---|
metastore → catalogs → schemas → objects | • Top-level container for Unity Catalog metadata and permissions • one per region per account. | |
sales_catalog.analytics.orders | • First namespace level that groups schemas • maps to environments, domains, or teams. | |
sales_catalog.analytics | Second namespace level (also called database) that groups tables, views, functions, volumes, and models. | |
SELECT * FROM sales_catalog.analytics.orders; | • Structured dataset registered in a schema • can be managed, external, or foreign. | |
SELECT * FROM sales_catalog.analytics.active_customers; | Read-only named query that executes at read time. | |
SELECT * FROM sales_catalog.analytics.daily_revenue; | Precomputed view whose results are stored and refreshed incrementally. |