A sudden slowdown usually means that more work is happening somewhere in the refresh path—or that you are timing a different kind of refresh than before. Start by separating Power Query Editor preview from a loaded refresh, then compare the query steps, source response, privacy settings and background activity that changed around the time the delay began.
First, identify which refresh is slow
Power Query can do work for an editor preview, a Power BI Desktop model refresh, an Excel query refresh, or a Power BI service refresh. Those are not interchangeable measurements. A slow preview does not by itself show that loading the model or workbook is equally slow, and timings from Desktop do not establish what is happening in the service.
As an Amazon Associate I earn from qualifying purchases.
Record the exact action you take and how long it takes. Compare it with the same action in the same file and environment, and note whether the delay is before data begins appearing, during transformation, or after source retrieval. Then ask what changed immediately before the slowdown: a new or reordered step, a larger source range, another source added to a merge, a privacy or credential change, an application update, or a different network or gateway route. These are useful leads, not proof of a cause.
Check whether a query step stopped folding
For supported structured sources, query folding lets Power Query translate transformations into work performed by the source. Microsoft defines it this way: A step folds when Power Query translates its transformations into a data source query and runs them at the data source.
(Microsoft Learn: Query Folding Indicators in Power Query.) If a step no longer folds, some work that previously ran at the source may instead be evaluated by the Power Query mashup engine.
#1 Best Overall
- The Microsoft Office 365 Bible: The Most Updated and Complete Guide to Excel, Word, PowerPoint, Outlook, OneNote, OneDrive, Teams, Access, and Publisher from Beginners to Advanced
- ABIS BOOK
In Power Query Editor, inspect the folding indicators at the last step that refreshed quickly and the first step associated with the slowdown. If the latter indicates that folding is not occurring, transformations up to that point are not being fully translated to the source. The indicator is a clue about evaluation, not a guarantee of total refresh time: folding support depends on the connector, source and individual transformations.
- In the affected query, select the last known-fast step and check its folding indicator.
- Select the first step after it that appears to change the behavior. Compare the indicators and, where available, inspect the query plan.
- Use View Native Query when the connector exposes it to see whether the step can be translated. If it is unavailable, that alone does not prove the entire query cannot fold.
- Review whether a recent transformation or step order changed what the connector can translate. Test a targeted change and compare timings using the same refresh path.
Folding is not limited to DirectQuery: Microsoft notes that For an Import table, it may be possible to achieve query folding.
Its value depends on the source and transformation, so do not assume every Import query should fold or that one non-folding step explains all elapsed time. See Microsoft’s query folding guidance for Power BI Desktop.
Why folding can make a large difference
Microsoft reports an example using the Power BI dataflows refresh mechanic with a dedicated Azure Synapse Analytics environment at DW2000c. In that specific example, average timings were 361 seconds without folding, 184 seconds with partial folding and 31 seconds with full folding. Those figures illustrate what folding changed in that setup; they are not a general benchmark or a prediction for another connector, source or workload. Details are in Microsoft’s query-folding examples.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Use diagnostics to locate the work
When folding indicators do not settle the question, Query Diagnostics can help show what Power Query evaluates and which requests it sends. Its coverage and behavior depend on the host application and diagnostic scenario; use it as evidence about the measured operation, not as a universal explanation for every refresh. Microsoft documents the feature in Query Diagnostics.
Where possible, compare diagnostic events with source-side telemetry or logs. That helps distinguish a slow source request from work taking place in Power Query. Keep the tested action consistent: diagnosing an editor preview does not automatically describe a Desktop model refresh or a service refresh.
Do not mistake multiple requests for repeated full loads
A refresh or preview can produce several source requests for different purposes. Power Query may retrieve schema, perform privacy analysis, download previews, profile columns or analyze folding. Seeing multiple requests therefore does not, on its own, establish that the final data was loaded repeatedly. Microsoft discusses these evaluations in Why does my query run multiple times?
Rank #4
If requests seem excessive, use diagnostics to classify them before trying to suppress them. In Power BI Desktop, background preview behavior can add work, especially in files with many queries. Distinguish that editor activity from the actual loaded-data refresh; Microsoft describes this behavior in its guidance on multiple evaluations and preview caching.
Free tools Windows power users keep installed
One-click scans. No signup required.
Review privacy isolation when queries combine sources
When a query combines sources, privacy levels can affect how Power Query evaluates the combination. Privacy isolation may require buffering, which can cost time. Check the privacy classifications and settings for the sources involved, and verify that they reflect the data’s sensitivity and sharing context.
Do not treat ignoring privacy levels as a routine speed fix. Microsoft warns that the setting can improve performance and functionality, but Power Query cannot ensure the privacy of data merged into the file or workbook. Changing it can expose confidential data. Make that choice only after a careful security review, and do not assume a Desktop privacy setting controls a service semantic model. See Microsoft’s Privacy Levels in Power Query and privacy levels in Power BI service.
If the query behavior looks unchanged, compare the source and route
If folding, diagnostics and privacy settings do not show a new pattern, check whether the source itself responds more slowly and whether the affected refresh uses a different network or gateway route. Compare the same query and refresh path in the affected environment; where available, source-side timing or telemetry can help isolate time spent outside Power Query.
A particular database, gateway, network, capacity, credential, application update or service incident cannot be identified from the slowdown alone. Establish those causes with evidence from the relevant source and environment rather than inferring them from elapsed time.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
