Forum Discussion
Combining different data sources/modelling with REST data
Hi mjc,
In Power BI, it supports to connect to the data source which use oAuth2 authentication, for example, SharePoint Online. So in your scenario, which kind of data source do you want to connect to in Power BI? Scheduled refresh is available in Power BI Service not in desktop.
As you mentioned "I created an app that pushes the data into Power BI using the REST API", do you mean Power BI rest API or another knid of API? If we connect to the data source in Power BI Desktop, we are able to perform modeling for data.
In your scenario, would you please share more information about your requirement, so we can understand it better?
Best Regards,
Qiuyun Yu
I have data in both freeagent and teamweek which I'd like to report on. Neither is supported by Power BI.
I understand that scheduled refresh is not available in Desktop, I meant that refresh worked in the desktop (and possibly on the initial publish) but scheduling fails.I push data into Power BI using the Power BI REST API, I'm not sure what other API might be available to push data into Power BI.
This data set is only available in the service, it cannot be used in he desktop so there is no method that I can see to combine it with other data.
Does that make sense?- I have no idea on what are freeagent and teamweek and what kind of data you'd like to report on, however Google tells that they both provide corresponding APIs, by which you may connect to them in Power BI Desktop via "get data"->web.
- As far as I know, you could push data via Power BI REST API, however there's no way to model and combine the dataset with other data.
- Alternatively, as to push data via Power BI REST API, since you've got the freeagent and teamweek data in some approach outside Power BI, so maybe integrate the other data with the forementioned locally and push the integrated data to Power BI Service.
Thanks but I'd take issue with your first suggestion - you cannot properly add a data source with get data->web that uses oauth2 authentication. You might be able to hard code a token in the headers but the whole point is that the tokens have a limited lifespan and require to be refreshed. There is no means to submit a refresh token to obtain a new access token and then subsequently use that (until it requires refreshed). It is possible to create function in M (or Power Query or whatever the language is called this week) and that is quite usable in Power BI desktop but it will not work in the Power BI Service so scheduled refresh is prohibited. Publishing a static dataset is definitely less then useful.
I'd be more than happy if you were correct and it was as simple as you suggest but it just isn't.
It really seems that my intial impression is correct. Power BI has an extremely powerful feature rich toolset for modelling data but it is in many instances of no practical use in the real world. My apps are just two examples that use OAuth2 for authentication, there are many others and many more that use other authentication methods, I'm far from alone in this.
Modelling the data outside of Power BI works but the question that then has to be asked is what does using Power BI actually buy me? Given the license restrictions and cost of Pro licenses would I be better off finding something else that does what I need?
I think Power BI shows massive potential but in my opinion it is still too immature to fill many of the claims being made about it. Maybe I'm just being naive in hoping that a product can meet the expectations being set by marketing.
