How to Edit Dataset in Power BI Workspace

Nothing's more frustrating than spotting a mistake in your dataset after you've already published it to a Power BI workspace. Whether it's a misnamed column, an incorrect data type, or the need to add a whole new calculation, you need a fix without breaking everything. This article walks you through the different ways to edit a dataset directly in the Power BI service for quick adjustments and covers the complete, reliable method of updating it using Power BI Desktop.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Before We Start: Dataset vs. Report

First, it’s important to understand a key distinction in Power BI: the difference between a dataset and a report. This concept is the source of a lot of confusion, but it's pretty simple when you break it down.

• A Dataset is the engine of your report. It contains your tables, relationships, data transformations from Power Query, and DAX measures. It's the structured data model you built that feeds all your visuals.
• A Report is the collection of visual pages - your bar charts, tables, line graphs, and slicers - that are built on top of the dataset.

Think of it like building a house. The dataset is the foundation, a plan, and the building materials. The report is the finished house with the decorated rooms and furniture. You can easily change the paint color or rearrange furniture (editing the report), but making changes to the foundation (editing the dataset) is a more involved process. A single dataset can be the foundation for many different reports, so changing it can have a wide-reaching impact.

Quick Edits: What You Can Do in the Power BI Service

In the past, you couldn't do much with a dataset once it was in the cloud. Thankfully, Microsoft has added some limited editing capabilities directly within the Power BI workspace. These are perfect for small adjustments that don’t require changes to the underlying data structure itself.

To access these features, navigate to your workspace, find your dataset (look for the orange dataset icon), click the three dots (...), and select "Open data model."

Here’s what you can do:

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

1. Create New DAX Measures

Need a quick calculation? You can write new DAX measures on the fly. Let's say you published your sales data but forgot to create a measure for total revenue.

• In the data model view, click the "New measure" button in the ribbon.
• A formula bar will appear. You can type in your DAX formula. For example:

Total Revenue = SUM(Sales[SaleAmount])

The new measure will appear in your Fields list and will be available to all reports connected to that dataset. You can’t, however, edit existing measures - this is strictly for adding new ones.

2. Edit and Create Relationships

If you forgot to link two tables or linked them on the wrong columns, you can fix that in the service. In the data model view, you'll see a diagram of all your tables and their relationships.

• To create a new relationship, simply drag a field from one table and drop it onto the corresponding field in another table.
• To edit an existing relationship, double-click the line connecting two tables to open the relationship editor. Here, you can change the cardinality (one-to-one, one-to-many) or the cross-filter direction.

3. Manage Row-Level Security (RLS)

Row-Level Security lets you control who sees what data. While you need to define the roles in Power BI Desktop initially, you can manage the user assignments for those roles in the Power BI service.

• Go to your workspace and find your dataset again.
• Click the three dots (...) next to it and select "Security."
• From there, you can select a role and add users or groups to it, defining what data they can access.

4. Update Data Source Credentials

If the password for your SQL database changes or you need to update an API key, this is where you do it.

• Click the three dots (...) next to your dataset and select "Settings."
• Expand the “Data source credentials” section.
• You'll see a link to "Edit credentials." Click it to enter the new username, password, token, or key required to refresh your data.

What You Can't Do in the Service

The in-service editor is great, but its limitations are important to recognize. You cannot use it to:

• Access the Power Query Editor.
• Add a new data source to your model.
• Transform data (e.g., split columns, unpivot data, change data types, add a conditional column).
• Rename existing tables or columns.
• Edit existing DAX measures or calculated columns.

For any of these more fundamental changes, you'll need to go back to the source: Power BI Desktop.

The Complete Guide: Editing a Dataset with Power BI Desktop

This is the standard, most reliable workflow for making significant changes to a dataset. It involves downloading a copy of a report connected to the dataset, making edits locally in Power BI Desktop, and re-publishing it to the workspace to overwrite the old version.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Step 1: Download the PBIX File from the Workspace

First, you need the original project file. Find a report in your Power BI workspace that uses the dataset you want to edit.

• Open the report.
• Go to File > Download this file.
• A dialog box will appear. Make sure you choose "A copy of your report and data (.pbix)." This downloads the file containing both the report visuals and the underlying dataset.

Note: If this option is grayed out, your Power BI admin may have disabled it. You may need to ask them to enable the "Download reports" tenant setting.

Step 2: Make Your Edits in PBI Desktop

Open the downloaded .pbix file in Power BI Desktop. Now, you have full access to all the editing tools.

Using the Power Query Editor

For changes to your data sources, columns, or data structure, you'll use the Power Query Editor. Click the "Transform data" button on the Home ribbon.

In the Power Query Editor, you can make almost any change you need, such as:

• Adding a new data source: Click "New Source" and connect to your data, then merge or append it with your existing tables.
• Splitting columns: Right-click a column header, choose "Split Column," and select your delimiter.
• Changing data types: Click the icon next to a column header (e.g., ABC for text, 123 for number) to change its type.
• Renaming a column or a query: Just double-click the name in the header or the Queries pane and type a new one.
• Adding a conditional column: Go to the "Add Column" ribbon and select "Conditional Column." This allows you to create new columns based on IF-THEN-ELSE logic — for example, categorize orders as "Large" or "Small" based on the order amount.

Using the Data Model View

Once you're done in Power Query, click "Close & Apply." Now you can go to the Model view in Power BI Desktop (the lefthand icon resembling a table diagram). Here you can make more extensive modeling changes than you can in the service, such as editing existing DAX columns, measures, or hierarchies.

Step 3: Save Your Changes

Before you publish, save your updated .pbix file to your local machine. It’s always good practice to keep a local copy as a backup.

Step 4: Re-publish to the Power BI Workspace

This is the final step. From Power BI Desktop, click the "Publish" button on the Home ribbon.

• Select the same workspace that the original dataset was in.
• Power BI will detect that a dataset with the same name already exists and will ask you if you want to replace it. Click "Replace."

This action will overwrite the old dataset in the Power BI service with your updated version. Any reports connected to this dataset will now reflect your changes.

Critical Best Practices for Editing Datasets

Overwriting datasets can be risky if not done carefully. Here are some pro tips to avoid common pitfalls.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Check for Downstream Dependencies

Before you publish, always check what other reports, dashboards, or apps are using the dataset. A renamed column, for example, will break any visual that uses it.

To see all connected items, go to your workspace, find your dataset, click the three dots, and select "View lineage." This shows you a complete dependency map so you know exactly what will be impacted by your changes.

Communicate with Your Team

If the dataset is used by multiple team members, let them know before you are about to push new changes. This prevents confusion when a visual suddenly "breaks" or data changes unexpectedly.

Keep Local Backup Files

Before overwriting a dataset, it’s wise to download the current version as a backup. While Power BI keeps some version history, having your own .pbix backups, maybe managed through a SharePoint or OneDrive library, is the safest bet in case you need to revert your changes quickly.

Final Thoughts

In summary, you can handle minor edits like adding measures or managing relationships right in the Power BI service for quick fixes. For anything more substantial, the best-practice workflow is to download the PBIX file, use the full power of Power BI Desktop to make your changes, and then re-publish it to the workspace.

While this process in Power BI is powerful, we know it can feel cumbersome, especially for team members who aren't data experts. This is why we built Graphed. We connect directly to your data sources and turn the process of building and modifying reports into a simple conversation. Instead of learning Power Query or worrying about overwriting a dataset, you can just ask in plain English — "add a filter for sales in the UK" or "change this bar chart to a line chart" — and get an instantly updated dashboard in seconds, not hours.