If you need to edit Fabric semantic model objects correctly, it is essential to understand how Microsoft Fabric and Power BI Desktop work together. Many developers connect to a model and assume they are modifying the shared dataset, when in reality they are only creating report-level changes. In this guide, you will learn how to properly edit Fabric semantic model structures using Power BI Desktop, including the difference between Connect mode and Live Edit mode, and when each option should be used for enterprise development.
In this guide, we will cover:
- How to create a semantic model from a Fabric lakehouse
- The difference between modeling in the browser and Desktop
- The critical difference between Connect and Edit modes
- How Live Edit works
- When Desktop is required for preview features
- Best practices for maintaining Fabric semantic models
For structured, hands-on Power BI and Fabric training
Why Editing Fabric Semantic Models Requires Strategy
Creating a semantic model in Fabric is simple. Maintaining it correctly is where many teams struggle.
Key considerations include:
- Where should core DAX measures live?
- Should changes be report-specific or centralized?
- When should you use the browser modeling experience?
- When is Desktop required?
- How do preview features affect your workflow?
Understanding these differences prevents fragmented logic and duplicated calculations.
Creating a Semantic Model from a Fabric Lakehouse
If you want to follow along, you can use the sample Retail Data Model from Worldwide Importers available inside Fabric.
To create a semantic model:
- Open your Fabric-enabled workspace.
- Navigate to your Lakehouse.
- Select New Semantic Model.
- Choose the tables you want to include.
- Confirm and create the model.
Fabric immediately generates a Direct Lake semantic model stored in the service.
At this stage, you can begin modeling in the browser.
Modeling in the Fabric Browser
The Fabric web modeling experience allows you to:
- Create DAX measures
- Define relationships
- Hide columns
- Adjust formatting
- Configure calculation groups
- Modify linguistic schema
For many workflows, this browser-based modeling is sufficient.
However, there are limitations:
- Calculated columns are not available in Direct Lake semantic models.
- Some preview features are not exposed in the service.
- Browser refreshes can interrupt complex DAX editing.
For advanced development scenarios, Desktop becomes essential.
Connecting to a Fabric Semantic Model in Power BI Desktop
When you open Power BI Desktop and connect to a Fabric semantic model, you have two options:
- Connect
- Edit
This distinction is critical.
The Connect Option: Report-Only Measures
If you choose Connect, you are connecting in DirectQuery/Live connection mode.
You can:
- Build visuals
- Create report-specific measures
- Add calculated columns at the report level
However, any measures you create here:
- Exist only in that report
- Are not saved to the shared semantic model
- Are not visible to other report developers
For example, if you create:
Total Sales = SUM(FactSales[Total Including Tax])
That measure is available only in your report file.
If someone else connects to the same semantic model, they will not see it.
This is useful for:
- Highly specific report logic
- Experimental measures
- Temporary analysis
But it is not appropriate for shared enterprise measures.
The Edit Option: Live Editing the Shared Model
If you choose Edit instead of Connect, Desktop enters Live Edit mode.
You will notice:
- No report view
- No data view
- Only modeling view
- A green banner indicating live editing
In this mode:
- You are editing the actual semantic model in Fabric
- Changes sync automatically to the service
- All users connecting to the model see your updates
If you create the same Total Sales measure in Edit mode:
- It appears immediately in the Fabric service
- It becomes available to all reports
- It becomes part of the centralized model
This is how shared enterprise measures should be created.
Live Edit Mode in Direct Lake
When working with Direct Lake semantic models:
- There is no imported data in Desktop
- The model connects directly to the Lakehouse
- All structural changes sync automatically
You can:
- Add relationships
- Create measures
- Modify formatting
- Configure calculation groups
- Adjust model properties
Everything updates in real time in the Fabric service.
This creates a seamless development experience while maintaining centralized governance.
When Power BI Desktop Is Required
Some features are only available in Desktop, especially during preview phases.
One key example is user-defined functions in DAX.
At certain stages of preview:
- The Fabric service may not allow function creation
- Desktop allows enabling the preview feature
- Functions can be created and deployed to the shared model
This means Desktop is often required to leverage cutting-edge capabilities before they become generally available.
Preview features frequently appear in Desktop first.
If your organization wants to experiment with advanced modeling features, Desktop is usually the entry point.
Common Mistake: Confusing Connect with Edit
One of the most common issues developers encounter is assuming that creating a measure while connected means they have updated the shared model.
They have not.
If you do not explicitly choose Edit mode, your changes are local to the report.
This can result in:
- Duplicate measures across reports
- Inconsistent logic
- Confusion among developers
- Governance challenges
Always confirm which mode you are using before building shared logic.
Pros and Cons: Browser vs Desktop
Browser Modeling
Pros:
- No installation required
- Quick access
- Full Fabric integration
Cons:
- Limited preview feature access
- Vulnerable to accidental browser refresh
- Slightly less responsive for heavy modeling work
Desktop Live Edit
Pros:
- Full modeling control
- Access to preview features
- Safer for long DAX development
- More tactile development experience
Cons:
- Requires installation
- Requires understanding of Connect vs Edit distinction
Best Practices for Maintaining Fabric Semantic Models
Centralize Core Measures
Create enterprise measures in Live Edit mode so all reports benefit from consistent logic.
Avoid Measure Duplication
Do not allow each report developer to recreate foundational metrics.
Use Desktop for Complex Modeling
For advanced DAX, preview features, and structural changes, Desktop provides greater control.
Clearly Define Ownership
Assign responsibility for maintaining the shared semantic model to specific developers or BI engineers.
Test Changes Carefully
Remember that Live Edit mode affects all connected reports immediately.
Validate:
- Existing visuals
- Dependent measures
- Performance
- Security
Final Thoughts
Editing semantic models in Microsoft Fabric requires understanding how Desktop and the Fabric service interact.
The key takeaway is simple:
- Use Connect for report-specific development.
- Use Edit for centralized model updates.
By mastering Live Edit mode in Power BI Desktop, you can maintain clean, scalable, enterprise-grade semantic models inside Fabric.
If you want deeper hands-on guidance on Power BI, Fabric, and advanced modeling techniques, explore professional training
