Incremental Refresh - Missing data in partition

[MASB91]

Hey - I have an incremental refresh policy set up on a dataset to refresh daily when a new day's worth of data is received. This refresh works fine and has been building up each day's worth of data with no issues for a couple of months.

However, last month when I republished the dataset to Power BI online (I made some changes on the desktop version), and refreshed the data on Power BI online to recreate the archive of daily data, the incremental refresh lost 28 days of data but did manage to archive a number of recent days (I think it was around 10 days in total).

As a workaround I added the 28 missing days to a historic excel file and appended it to a table. However, I have since republished the dataset to Power BI Online again after making some more changes to the desktop version and when I have refreshed this to create the archive, 30 different days of data is now missing but the latest 8 days have been incrementally refreshed/archived.

Does anyone have any experience of incremental refreshes losing partial data? If so, can anyone help me with this issue?

[cmaier]

I am having a similar issue, was a solution found?

[bcdobbs]

Can you send a screen shot of the configuration screen?

[MASB91]

This is the configuration. The data comes from excel files which are saved in a document library in SharePoint. The table then filters (in PowerQuery) on the date the file was created using the parameters "After or Equal to RangeStart" and "Before RangeEnd". There are 8 days worth of data in each file, and I wish to incrementally refresh the first day of this data (there is an additional filter in the power query steps for this).

The parameters set up for this table are RangeStart "01/07/2022" and RangeEnd "02/07/2022". The missing data covers the period 24/06/2022 to 23/07/2022. The incremental refresh has partitioned the data in PBI online from 24/07/2022 to present.

[bcdobbs]

Is it loosing the oldest 30 days?

[MASB91]

Yes, it is losing the oldest 30 days and only bringing in a small number of the most recent days.