OneLake Integration Power BI: Connect Semantic Models to Fabric

Do you have existing Power BI semantic models and want to use them within Microsoft Fabric? This guide walks you through how to enable OneLake integration in Power BI and connect your semantic model data to a Fabric Lakehouse, without needing to rebuild your models from scratch.

What Is OneLake Integration?

OneLake integration allows you to expose Power BI semantic model data as Delta tables inside Microsoft Fabric. While it doesn’t convert semantic models into Fabric-native formats, it lets you connect and query that data using Lakehouse features. This is particularly useful if you’re already working in a Fabric-enabled workspace and want to unify your analytics environment.

Key Requirements

Before you start:

• Your semantic model must use Import Mode (DirectQuery is not supported).
• You need a Fabric-enabled workspace (Fabric capacity or free trial).
• The report and model should already be published to that workspace.

Step 1: Publish Your Power BI Report to Fabric

1. Save your Power BI report that contains the semantic model.
2. Publish it to a Fabric-enabled workspace in the Power BI service.
3. Confirm that the workspace uses a Fabric capacity under Workspace Settings.

Step 2: Enable OneLake Integration

Once published:

1. Go to your semantic model in the Fabric workspace.
2. Select the ellipsis (…) > Settings.
3. Under OneLake Integration, toggle it on and click Apply.

Step 3: Refresh the Semantic Model

This step is critical:

• Refresh the semantic model manually or via schedule.
• This process converts imported tables into Delta tables, making them accessible in OneLake.

Step 4: Connect from the Lakehouse

Inside your Fabric Lakehouse:

1. Click the ellipsis (…) > Create Shortcut.
2. Choose Microsoft OneLake and locate your published semantic model.
3. Select the tables you want (e.g., Product, SalesOrderDetail).
4. Rename and organize them as needed and create the shortcut.

Now these Delta tables are part of your Lakehouse and available for querying or integration across Fabric.

How It Works Behind the Scenes

• Power BI semantic model tables in Import Mode are exported as Delta Parquet tables.
• These are stored in OneLake and refreshed alongside your report.
• The data is readable via Lakehouse, but the Power BI report continues to function as usual.

When (and When Not) to Use It

This integration is best when:

• You want to leverage Power BI data inside Fabric without duplicating ETL pipelines.
• You need a lightweight data sharing mechanism between Power BI and Lakehouse.

Avoid using it:

• As a wholesale replacement for structured ETL or data warehouse logic.
• If your model is in DirectQuery or needs complex real-time access.

Learn More About OneLake and Fabric

New to OneLake or Lakehouse concepts? Visit our on-demand training at Data Bear Power BI Training to explore Fabric fundamentals, Lakehouse architecture, and best practices.

Final Thoughts

With OneLake integration, Power BI and Fabric work together more seamlessly than ever. By exporting semantic model tables into Delta format, you enable reuse of analytical data across tools without losing the original Power BI reporting experience.