[Unable to combine data] accessing data sources that have privacy levels

AnonymousAnonymousGilbertQ

Sorry to reply to this so late. I am consuming several RESTful API's, so this ended up being a blocking issue for me. I spoke with Power BI support about the issue and there a few scenarios.

The first is "https://www.domain.com/rest/api" or similar. If you try authenticating against this URL alone, you'll usually find yourself unsuccessful or redirected. The way the Power BI service online was explained to me, is that it tries to run through and authenticate the endpoints in your dataset before allowing/attempting a scheduled refresh. In hindsight, this makes sense because your desktop application can implicitly be trusted by being your physical property, where the Power BI service on Microsoft's servers needs an extra step to verify access. Doing authentication this way also avoids tying up resources that are going fail anyway.

The second is "https://www.domain.com" with a parameter passed and appended to the baseurl. This fails for the same reason, but the difference is the Power BI service is set up to inform you that your datasource doesn't support scheduled refresh before it ever attemtps to authenticate against the endpoint. This is to avoid you putting in any arbitrary parameter that could potentially cause issues, therefore tying up resources. It would also take some work to "build" the URL's, if you will, in order to authenticate them ahead of time and support parameterized baseurls.

The last scenario relevant to this topic is "https://www.domain.com" and the use of the RelativePath property defined with a variable in the options list parameter. If you've tried this, you know that your dataset in the Power BI service will have no indication of an issue until it tries to refresh. It fails for the same reasons above, but Power BI doesn't have a check built in like it does when appending a parameter to the baseurl.  For those who mentioned using JIRA data, I know that API key auth support is coming down the pipe on their end sometime soon and the current content pack implementation (https://www.domain.com/jira/rest/api/2/search) works with scheduled refresh today. If you are trying to access different resource perspectives, one of the three above scenarios should describe any issues you are having (I'm consuming the remotelinks endpoint which blocks scheduled refresh for me)

At the time, support told me that supporting scheduled refreshing for parameterized Web datasources was not on their roadmap given the large changes they have made this year and are continuing to make concerning Power BI. That led me to look into setting up a gateway, but I don't believe the scheduled refresh there differs from the Power BI service concerning the Web datasource (please correct me if I'm wrong). My current solution is refreshing my reports manually as often as I need which is not really a workaround, but I wanted to share my experience so far just in case.

One suggestion that seemed promising was to store the report in OneDrive for Business locally for automatic syncing from desktop, but the only repo to interact with Power BI desktop is using a UI testing framework and hasn't been updated in awhile. I was hoping to find a useful set of cmdlets for Powershell to interact with Power BI desktop or some other way to refresh a report automatically to make this a good workaround, but haven't found anything so far. The last option that GilbertQ mentioned is creating a custom connector for the datasource you are working with. This currently does not support OAuth and I've heard a handful of other scenarios where consuming an API that you don't own may not be supported. I'm dealing with about a 50/50 split as far as the API's I'm consuming, so I decided not to commit considering I wouldn't be gaining scheduled refresh anyway.

Hopefully this is helpful. I am continuously investigating solutions to this problem, so I wll try to update this thread if I find any other potential routes.