Power BI Displaying Counts differently in Service vs Desktop

[wonka1234]

Hi,

After publishing my report to service i am getting a difference in counts in a chart.

If i export data from the chart in service and desktop I am getting

Desktop - 

Report Date	Count of People	People
3/1/2025 0:00	57	A
3/1/2025 0:00	15	B
4/1/2025 0:00	58	A
4/1/2025 0:00	11	B
5/1/2025 0:00	56	A
5/1/2025 0:00	13	B

PBI Service -

Report Date	Count of People	People
3/1/2025	57	A
3/1/2025	15	B
4/1/2025	59	A
4/1/2025	11	B
5/1/2025	57	A
5/1/2025	13	B

Whats going on here? This never occured before. Shouldnt PBI service show exactly what I did in my PBI desktop?

[wonka1234 · Accepted solution]

thanks all for the replies. I clicked "reset filters slicers and other data view changes you've made" and it worked. Why does this work? Kind of confused with it.

It seems like if i want accurate results i have to "reset" them each time I open the dashboard.

I have disabled it in REport Settings -> Check Dont Allow Users To Save filters on this file for PBI Service

It almost seems like a particular view is saved when I publish to PBI service and have to reset to default each time.

Could this be related to bookmarks?

After further investigation after unchecking "data" in my bookmarks (used for navigation) it seems to work. But why? I dont understand.

[wonka1234 · Accepted solution]

thanks all for the replies. I clicked "reset filters slicers and other data view changes you've made" and it worked. Why does this work? Kind of confused with it.

It seems like if i want accurate results i have to "reset" them each time I open the dashboard.

I have disabled it in REport Settings -> Check Dont Allow Users To Save filters on this file for PBI Service

It almost seems like a particular view is saved when I publish to PBI service and have to reset to default each time.

Could this be related to bookmarks?

After further investigation after unchecking "data" in my bookmarks (used for navigation) it seems to work. But why? I dont understand.

[v-sgandrathi · Accepted solution]

Hi @wonka1234,

Thank you for the update!

The issue you observed is due to bookmarks saving filter/slicer states when the “Data” option is checked. Even if filters are not visibly changed, bookmarks can silently apply these saved states, resulting in differences in visuals.

Also, Power BI Service may retain previous filter states even after changing settings such as "Don’t allow users to save filters." Clicking "Reset to default" clears these saved views and loads the original published state.

Please follow these steps to resolve the issue:

1. Uncheck “Data” in bookmarks used for navigation.

2. Republish the report after verifying slicers are in their default state.

3. Advise users to reset the view if visuals appear incorrect.

If my response was helpful, consider clicking "Accept as Solution" and give us "Kudos" so that other community members can find it easily. Let me know if you need any more assistance!

Regards,
CST_Microsoft Fabric Community.

[wonka1234]

Whats strange is that i can still reset filters after disabling it. Or is it different because I am a member of the workspace? If i hover over it, it does say the author disabled filters, but i can still click it.

[v-sgandrathi]

Hi @wonka1234,

Even when the "Don't allow users to save filters" setting is enabled, workspace members (especially Admins or Members) can still interact with filters and use the "Reset to default" option. This is because Power BI assumes that authors and contributors may need to test or troubleshoot different states of the report.

The tooltip indicating that filter changes are disabled by the author is accurate for viewers but not enforced for workspace members.

Reset to default still works for you because Power BI retains a personal view until it's explicitly reset, even if saving new filter states is disabled.

To fully test the experience, try viewing the report in a private browser or as a user with the Viewer role.

Troubleshoot visualizations in Power BI - Power BI | Microsoft Learn

Thank you.

[ajaybabuinturi]

HI @wonka1234,

I would like to let you know that this should be Date Granularity Mismatch

Desktop date column → shows full datetime (3/1/2025 0:00)

Service date column → shows just date (3/1/2025)

This indicates a data type formatting difference, and possibly a grouping difference in aggregation.

Ensure your Report Date column is explicitly converted to Date (not datetime)

ReportDate = DATE(YEAR([OriginalDate]), MONTH([OriginalDate]), DAY([OriginalDate]))

Or in Power Query:

Use Transform > Data Type > Date (not Date/Time).

Rebuild visuals using this cleaned date column to avoid accidental mismatches due to time values.

Thanks,
If you found this solution helpful, please consider giving it a Like👍 and marking it as Accepted Solution✔. This helps improve visibility for others who may be encountering/facing same questions/issues.

[wonka1234]

@ajaybabuinturi  tried it, it was already in "date" format and not "datetime". I also tried putting it at the end of my power query steps and still same result 😞

[v-sgandrathi]

Hi @wonka1234,

Thank you for confirming that your Report Date column is in Date format and that placing the transformation at the end of Power Query did not resolve the mismatch. Here are the steps you should check:

Use Performance Analyzer in Desktop to capture the exact DAX query being run and compare it with the Service version.
In Desktop, view Performance Analyzer, start recording, and refresh the visual. Copy the DAX query and test it in DAX Studio or Tabular Editor.
Even if the column is in Date format, locale differences can affect date interpretation. In Power BI Desktop, go to File, then Options, and check Regional Settings. In Power BI Service, review the workspace or tenant-level locale settings.
There might be subtle differences in the underlying data between Desktop and Service due to refresh timing or source configuration.
Sometimes, relationships between tables, especially on date fields, behave differently in Service due to refresh order or model inconsistencies. Open Model View and ensure the relationship between your fact table and date table is active. The join should be single-directional unless needed otherwise, and the cardinality should be correct.

I'm attaching the resolved thread that should assist you in addressing the issue:
Solved: power bi desktop shows different data than in powe... - Microsoft Fabric Community
Solved: Data shows differently after publishing to service - Microsoft Fabric Community

Please Accept as solution if this meets your needs and a Kudos would be appreciated.

Thank you and we're happy to assist you on the Microsoft Fabric Community Forum.