Announcing support for Materialized views and streaming tables in Mirrored Azure Databricks catalog

Microsoft Employee

9 hours ago

Access more Unity Catalog objects through the existing mirrored catalog experience in Microsoft Fabric

We are expanding Mirrored Azure Databricks catalogs in Microsoft Fabric to support materialized views and streaming tables alongside tables.

These objects remain created and managed in Azure Databricks, while Fabric provides read-only access to the underlying data through shortcuts. This brings more of your Unity Catalog environment into Fabric without copying or replicating the data.

What's new

A Mirrored Azure Databricks catalog can now include support:

Tables
Materialized views
Streaming tables

When configuring the mirrored catalog, select the Unity Catalog catalogs, schemas, and supported objects you want to make available in Fabric. Fabric mirrors the selected catalog structure and creates shortcuts to the underlying data.

This enhancement uses the existing Mirrored Azure Databricks catalog item and configuration workflow. There is no separate Fabric item type for materialized views or streaming tables.

Why this matters

Materialized views store precomputed results that can simplify recurring analytical queries. Streaming tables continuously process incoming data and maintain updated results for scenarios that require fresher information. Both object types can play an important role in the datasets that organizations publish for analytics and reporting.

Supporting these objects in a mirrored catalog gives Fabric users access to more of the curated data already available in Azure Databricks. It also allows platform teams to preserve clear ownership: Azure Databricks remains responsible for creating and processing the source objects, while Fabric provides experiences for discovering and analyzing the available data.

Prepare objects for external access

Before a materialized view or streaming table can appear in Fabric, configure it for external data access in Azure Databricks, complete the applicable prerequisites, and grant the required Unity Catalog permissions to the identity used by Fabric. Then refresh or synchronize the mirrored catalog.

Work with data in Fabric

The mirrored catalog includes an automatically generated SQL analytics endpoint for exploring and querying the available data with T-SQL. You can also create a Power BI semantic model and use Direct Lake mode for reporting.

Keep catalog changes in sync

For new mirrors, automatic synchronization of future catalog changes for the selected schema is enabled by default. Supported additions and deletions of schemas, tables, materialized views, and streaming tables can be reflected in Fabric when they meet the configured selection and external-access requirements.