Databricks ADBC connector for Power BI and Microsoft Fabric (Generally Available)

This release is part of the broader ODBC-to-ADBC transition for Power Query connectors described in the Transition from ODBC to ADBC drivers in Power BI and Microsoft Fabric documentation.

What’s new

Existing connections in your semantic models, dataflows, and paginated reports continue to work unchanged — the migration path is currently opt-in per connection, per workspace, or per tenant.

Highlights:

• Improved security - ADBC drivers incorporate security enhancements, such as memory safety and garbage collection.
• No changes to existing behavior - This transition moves away from the embedded ODBC drivers that ship with Power BI. It does not change behavior for the ODBC connector when you use a separately installed ODBC driver.
• Efficient large dataset support - ADBC provides a set of standard interfaces for interacting with Arrow data, which is especially efficient at fetching large datasets with minimal overhead and no serialization or copying.
• Supported authentication - Databricks ADBC supports OAuth (user), OAuth (M2M / Client Credentials), and Microsoft Entra ID.

Connectors covered by this release

How to try it

You can test the ADBC driver now. We suggest you start soon to help ensure the eventual deprecation of ODBC and migration to ADBC runs smoothly. There are three ways to begin validating ADBC, ordered by how much control you want.

1. Opt in per connection using Implementation="2.0"

The fastest way to compare ADBC and ODBC for a specific report or dataset. Add the Implementation="2.0" option to the source step in your Power Query expression:

Source = Databricks.Catalogs("adb-1234567890.1.azuredatabricks.net",
                             "/sql/1.0/warehouses/<warehouse-id>",
                             [Implementation="2.0"])

Refresh the dataset and compare row counts, column types, and totals against your production baseline. If you want to force a side-by-side comparison against ODBC, use Implementation="1.0".

2. Override the default for a validation workspace

If you want to validate several reports together, a workspace admin can turn ADBC on for a single workspace while the rest of the tenant continues to use the legacy driver. Publish the reports you want to test into that workspace and run your normal refresh schedule.

3. Enable ADBC by default at the tenant level

When you’re confident with the results, the tenant admin can enable Users can connect to data sources by using Apache Arrow database connectivity (ADBC) in the Admin portal. New connections across the tenant will use the ADBC driver by default, and Desktop and the service stay aligned on the same driver.

What to expect during validation

The ADBC driver is designed as a compatibility-preserving replacement for the Simba Spark ODBC driver. Validate your key reports in Desktop before enabling ADBC at the tenant level and report any regressions to [email protected].

Explore Transition from ODBC to ADBC drivers in Power BI and Microsoft Fabric for more information.

What’s next

The Power BI service will move from ODBC to ADBC through the first quarter of 2027. We recommend that you start validating today, so the tenant-level move is a non-event when it lands.

We’d love your feedback

Try the new Databricks ADBC driver. If anything looks different from what you'd expect — a refresh that fails, a total that changed, a performance regression — send the tenant ID, connector, M expression, and what you observed to [email protected].

Get started this week:

• Read the Databricks connector documentation.
• Install the latest Power BI Desktop.
• Pick one production report on Databricks and add Implementation="2.0" to its source step.
• Compare row counts, column data types, and key totals against your production baseline.
• Report anything unexpected to [email protected].