The recommended approach to Power BI as an enterprise solution is to have your ‘shared data set in a pbix file where you can just create test visuals to check everything is ok.
Then create the real reports in a separate pbix file live connected to the dataset.
But what happens if you create great reports, including Bookmarks etc in the shared data set before deciding to move everything?
Changing the data source
Instead of creating a new pbix file and copying all the visuals in (this wont really work if you have bookmarks and more complex reporting set up)
First, save your ‘Shared pbix’ file with another name which will becomes the reporting pbix file. The original pbix file complete with the original reports can be left as is.
You could change the data sources over to the ones in the data set. However this creates an issue.
The Data set pbix file sits on dataflows.
If you currently try and change to data sets, Data sets aren’t offered.
So what do you do?
In this example all the M Code for the Query should be done in the Dataflow within service. the only M Code in the shared data set should be the connection to the data flow.
All the Queries should be deleted. If you make sure your dataflow queries have the name you want in the dataflow, Re adding them back will be fairly straight forward
Close and Apply your Query
Back in Power BI all your visuals show as having issues. There is no data.
Go to Get Data and now you can choose datasets
Find the data set and select. This should then load in Live data from the data set and your visuals will be displayed again.
Publish this back up to service.
At this point you can look at Lineage to view that you are using the power BI recommended approach
When I create the app for Viewing I simply hide the Dev reports from the users.
This appears to be the speediest way of moving your reports to their own file, if you have already done a lot of reporting work in the ‘shared data set’