Forum Discussion
Cannot Modify Scheduled Refresh or Create a Dataset
First, Ignore the error message that doesn't actually mention a source, it's a bug.
Datasets are best created with Power BI Desktop, forget about the Get Data menu in the service as it's extremely limited. When you publish from Power BI Desktop to the service, that's where the data gateway will be used. The Desktop app is for data modeling and visualizations, it's not for scheduled refreshes, so the gateway is irrelevant to the desktop app.
First, Ignore the error message that doesn't actually mention a source, it's a bug.
Datasets are best created with Power BI Desktop, forget about the Get Data menu in the service as it's extremely limited. When you publish from Power BI Desktop to the service, that's where the data gateway will be used. The Desktop app is for data modeling and visualizations, it's not for scheduled refreshes, so the gateway is irrelevant to the desktop app.
Thanks, that helps.
I created a new dataset in Power BI Desktop, pushed it to Power BI Service and then I can map it to the required Data Source. Since I did this, I can also edit the original dataset refresh schedule (go figure). While this bug is present, how are people editing referesh schedules?
Also, a bit of a scope creep question, what is the recommended way to change a Dataset/Data Source for a report? Without having to re-create all the visuals.
#1 so far when I've run into the bug it was transient so I could edit scheduled refreshes after a browser refresh.
#2, the recommended way would be to substitute the Source step in Power Query, which will work if the output of that step is exactly the same as the previous one (otherwise it's likely the rest of the query will break because of mismatched columns, data types etc). Obviously don't rename columns either or dependent measure/visuals will break.
If you have a bunch of dependent queries, it would be a good idea to use a parameter to make things more portable in the future. And since you'll have to edit your queries anyway, you might as well look into dataflows and do all your transformations there, upstream of your dataset.
Thanks, again, appreciate your time.