Zero-copy access to OneLake data in Azure Databricks (Preview)
Most data estates are not single platform, and that is not a problem. The challenge is what usually comes next: extra copies, extra pipelines, extra refresh schedules, and endless debates about which version is the truth.
Today, we are introducing OneLake catalog federation (Beta) in Azure Databricks Lakehouse Federation, which simplifies multi-engine analytics by enabling Unity Catalog in Azure Databricks to query data stored in OneLake. This allows you to analyze Fabric tables without copying the data.
Figure: Diagram Azure Databricks reading OneLake data
Why this matters
Fabric is built around a simple idea: your data should be usable the moment it lands, and it should stay governed and consistent wherever it is consumed. OneLake is the foundation that makes that real by giving your organization a single, shared data lake that is easy to manage, secure, and scale.
OneLake catalog federation extends that promise even further. It allows more teams to use the same curated data products in OneLake without creating additional copies or building parallel pipelines just to satisfy different tools. That means fewer moving parts, fewer refresh problems, and far less time spent reconciling datasets. You keep OneLake as the source of truth and the Fabric experience intact, while expanding how broadly those OneLake data products can be used across the organization.
What you can do
- Discover Fabric tables from Unity CatalogOnce connected, your Fabric schemas and tables show up in Unity Catalog through a foreign catalog that stays in sync.
- Query OneLake data from Databricks compute
You run Databricks SQL and notebooks as usual, using catalog.schema.table naming. - Keep OneLake as the source of truth
No extra storage copies. No extra refresh jobs.
Conceptual overview
- You create a OneLake connection in Unity Catalog.
- You create a foreign catalog that points to a specific Fabric item.
- Databricks syncs the metadata, so schemas and tables appear in Unity Catalog.
- Queries run on Databricks compute while reading the data in OneLake.
Follow the full walkthrough in the documentation: Enable OneLake catalog federation – Azure Databricks | Microsoft Learn
Beta limitations
This capability is in Beta. Explore the specific requirements and supported configurations detailed in the Beta limitations documentation.
Try it out
OneLake is about removing friction between data and value. OneLake catalog federation is another step in that direction: fewer copies, simpler architecture, and broader reuse of the data products you already build in Fabric.
A shared OneLake foundation unlocks new possibilities for what teams can build next.
