How to Discard Pending Changes in Power BI

Ever found yourself deep inside Power BI's Power Query Editor, making transformation after transformation, only to realize your data preview now looks completely broken? You've tried splitting columns, filtering rows, and merging queries, but instead of clarifying your data, you’ve made a mess. This guide will show you exactly how to discard those unwanted pending changes and get you back to a clean slate, no stress needed.

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

What Are "Pending Changes" in Power BI, Anyway?

To understand how to get rid of them, it helps to first know what "pending changes" are. When you open the Power Query Editor in Power BI (by clicking "Transform data"), you're entering a workshop where you prepare your data before it goes into the final report. Every action you take - removing a column, changing a data type, filtering a value - is recorded as a step in the "Applied Steps" pane.

Think of it like adding items to a grocery list. You can add milk, eggs, bread, and so on. But none of those items are actually in your fridge yet. In Power BI terms, these listed steps are your "pending changes." They are a recipe of instructions waiting to be executed.

The changes are only made permanent in your data model when you click the "Close & Apply" button. That’s the moment you leave the Power Query "workshop" and Power BI runs all those steps on your actual data. Until then, everything is provisional, and thankfully, very easy to undo.

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

A Common Scenario: The Experimental Data Cleaning Spiral

Let’s paint a picture that might feel familiar. As a marketer, you've just exported a raw CSV file of campaign-level performance from an ad platform. The data is messy and needs cleaning before you can build any useful visuals.

You open it in the Power Query Editor and start experimenting:

• You try to split the ‘Campaign Name’ column by a delimiter to extract the country code, but you choose the wrong delimiter.
• Then you attempt to unpivot a few columns to get your metrics into a cleaner format, but it results in a thousand extra rows that don't make sense.
• You filter out a few rows, but accidentally filter out essential baseline data.

Before you know it, the data preview tables look scrambled, and your "Applied Steps" list is a chain of questionable decisions. The last thing you want to do is hit "Close & Apply" and let this monster loose on your data model. So, how do you back out safely without closing Power BI and losing everything?

Step-by-Step: How to Discard All Pending Changes

When you're sure you want to scrap every change you’ve made since the last time you applied them, Power BI gives you a straightforward escape hatch. It's not hidden, but if you don't know where to look, it can feel like you're trapped.

The Main Methods for Discarding Changes

The simplest way to bail on your changes is by telling Power BI not to apply them when you try to leave the Power Query Editor.

1. With the Power Query Editor open, go to the top left corner and click the Close button in the ribbon (not Close & Apply). Or you can simply click the 'X' button in the top right of the window as if you were closing any other application window.
2. A dialogue box will appear, warning you: "You have queries that have not been applied." This is Power BI's safety net.
3. You will be presented with three primary choices:

For some versions of Power BI, this dialog may look slightly different, perhaps giving choices like "Apply," "Don't Apply," and "Cancel." In that case, clicking "Don't Apply" accomplishes the same goal as "Discard changes."

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

A Note on Undoing a Discard

It's important to understand that this action is final for your current session. Once you discard the changes, that specific sequence of unapplied steps is gone for good. There isn't an "undo discard" button. Of course, the Power Query Editor will open back up to the last correctly applied state, and you are free to try making those transformations again, but the batch of changes you just wiped cannot be instantly recovered.

A Smarter Approach: Selectively Removing Changes

What if it’s not an "all or nothing" situation? Often, you've made five great transformations and just one bad one that's messing everything up. In this case, you don't need to discard everything. You just need to perform some surgical removal.

This is where the Applied Steps pane on the right of the Power Query Editor becomes your best friend.

Here’s how to use it:

1. Look at the list under "Applied Steps." Every transformation you've made to the selected query appears here in chronological order.
2. Identify the step that caused the problem. For example, you might see "Split Column by Delimiter" or "Filtered Rows."
3. To remove that single step, simply click the small "X" icon to the left of the step's name.

That's it! Power Query will instantly remove that step and re-run the subsequent ones, updating your data preview. You can backtrack as many steps as you need this way, giving you precise control over your data cleaning workflow without having to throw out all your good work.

Using the Applied Steps list proactively is one of the most powerful habits you can develop in Power BI. It allows for fearless experimentation because you always know you are just one click away from reversing a mistake.

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

A Proactive Workflow to Avoid Reporting Accidents

Knowing you can discard or delete changes gives you a safety net, but a good workflow can help prevent you from needing it as often.

• Review as you go: After each significant transformation, glance at your data preview. Does it look right? Is the outcome what you expected? It’s much easier to catch and fix one bad step than twenty.
• Rename your steps: By default, steps get generic names like "Changed Type1." Right-click a step and choose "Rename" to give it a descriptive name like "Set Sales Column to Currency." This makes your work much easier to debug later.
• Duplicate queries for experiments: If you're about to try something complex or risky, right-click the query in the left-hand pane and choose "Duplicate." Now you can experiment safely on the copied version without touching your original, clean query.

Final Thoughts

Getting tangled in Power Query transformations is a common rite of passage, but it doesn't have to be a point of stress. By understanding how to discard all pending changes or selectively delete individual applied steps, you can explore and clean your data with full confidence, knowing you can always backtrack if a transformation goes wrong.

Building effective reports often involves navigating the complexities of tools like Power BI, from managing pending changes to mastering DAX formulas. This process can be powerful but also time-consuming. That's why we created Graphed. We simplify the entire analytics workflow by allowing you to connect your data sources - like Google Ads, Shopify, or Salesforce - and build real-time dashboards just by asking questions in plain English. There’s no need to worry about applied steps or manual data cleaning, just describe the report you need, and our AI does the rest, delivering actionable insights right away.