Refresh brought in incorrect data

[TakeoffGoatMigo]

Today I checked a report and the numbers were way off from what they should be.  I couldn't begin to understand what could have changed in the dataset to do this.  I checked source data, queries, did refreshes on desktop, etc.  Once the online service refreshed again, as it is scheduled to every 90 minutes, the data was back to normal.

Has anyone else experienced this?  I've seen data change DURING a refresh, but never as a result of one.  This inconsistent data as a result of refresh is becoming a big problem for our organization.  Any help or assistance on what to do is appreciated.

[Anonymous]

Hi @TakeoffGoatMigo ,

Based on the information you have provided, we understand that you are experiencing an intermittent error presenting incorrect data. Can you provide us with your version of Power Bi Desktop, the connection mode of your data source, and the time you discovered the issue so we can confirm if this is a known issue.

Best Regards,

Ada Wang

[TakeoffGoatMigo]

I am using Desktop version @2.124.1805.0 64-bit (December 2023) with a live connection to a published dataset to the PBI server.  The original dataset uses an import connection to its source.  I discovered this issue about 6 months ago and have experienced it multiple times since, including today.  This problem has spanned over multiple datasets, all with the same connection as I just described.  How would I contact a Microsoft rep to discuss further?

[Anonymous]

Hi @TakeoffGoatMigo ,
Unfortunately this does not appear to be a known issue, you can open a support ticket at: https://powerbi.microsoft.com/en-us/support/

Best Regards,

Ada Wang

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[edhans]

Is there anything going on with the source data during the refreshes? For example, could it be making snapshots at the time, or is someone rolling data from TEST to PROD and then refreshing and your semantic model is picking it up mid-update?

I work with hundreds of semantic models across many orgs of all kinds of sources and haven't seen where data is just incorrect. There is always an underlying cause.

[TakeoffGoatMigo]

There is no suspected activity related to the source data.  Everything is coming from CRM, which we checked during this, and the data looked fine.  All of it in the production environment with nothing being rolled over.  There have been times where the data changes during a refresh, but goes back once the refresh completes.