I have found another problem with the service. The first one was the use of DATESINPERIOD where more than about 400 records were present.
This one is about a column whose contents are strings that work fine on the desktop. They turn to blanks on the service.:
The blank fields are used by a Slicer I have that lets one select the Manager for a particular project. The field is found in a record which is expanded by an M language query during load.
On the desktop, after complex maneuvering of data, i end up with 138 records of financial data from the current month. I have a slicer that uses the field Project_x0020_Manager to filter data. Project_x0020_Manager does contain data.
If I publish to the service, the Project_x0020_Manager column has blanks in it. The slider therefore only has "[blank]" as a choice.
On the service, to verify the problem is not with the visualization, I temporarily add a visual filter. When I select the filter, the 138 instances of "blank" are reported.
To bring the problem full-circle, I then download the pbix from the service. When I load it on the desktop version, the column is indeed blank as it online.
For a final check, using the same downloaded file, I refresh all the data from the sources.
Once that is done, the data for that column reappears
The data in the column is sourced from a sharepoint list, but the column itself is expanded from a sharepoint record, thus:
#"Expanded FieldValuesAsText" = Table.ExpandRecordColumn(#"Changed Type2", "FieldValuesAsText", {"Project_x0020_Manager", "Author", "Editor"}, {"Project_x0020_Manager", "Author.1", "Editor.1"}),
Due to the sensitivty of the data, I can only share the file securely with you if you're a Microsoft employee.
-thanks
-e
