Unity Catalog
info
Unity Catalog (UC) provides a centralized governance solution for managing data and AI assets across Databricks environments.
More information on:
UC sits between your workspaces and data lake.

UC objects and the three namespace convention

--- catalog: first namespace.
--- schema: second namespace.
--- table / view / volume / model: third namespace
SELECT * from my_catalog.my_schema.my_table;
--- Example: Select the gold table sales from the galaxy project on the development catalog.
SELECT * from dev.galaxy_gold.sales;
-- Set a default catalog or schema
USE CATALOG dev;
USE SCHEMA galaxy_gold;
--- Uses the defined catalog and schema
SELECT * FROM sales;
UC Governs All Data, ML, and GenAI Interactions
When a user or service principal performs any of the following actions:
- Use an UC object.
- Creates an UC object.
- Populates a UC object (catalogs, tables, schemas).
- Deletes an UC object.
- Queries a UC object (query, view, table).
- Runs an UC object (function, model).
UC verifies the associated grants and permissions, as illustrated in the following diagram:

Lessons learned
What is Unity Catalog?
Is the centralized governance layer designed to manage data access, security, lineage, and governance across Databricks workspaces, enabling unified data management and secure collaboration.
Where is the data stored?
The data, metadata and AI models are always stored in your organization cloud account.
How should every data+AI interaction be done?
Through UC.
What should be avoided?
- Accessing data, models or GenAI agents using hardcoded credentials.
- Configure data access on the clusters using external libraries + environment variables.