Forum Discussion

Anonymous's avatar
Anonymous
Not applicable
2 years ago

connecting Multiple data sources, dataflow or dataset ?

Situation; I have two types of data; 1. we have multiple data sources pulling end user data from jira into excel files (3) then the ETL process is done in power bi desktop. 2. we also have end user data in excel files(6) and the ETL process is done in power BI desktop.
For both cases all the reports and datasets (5) get published to a power bi workspace, where more reports are then created (12 - 7 new, and 5 original).These reports are then published to an app and multiple audiences are then given access to certain reports.
RLS is placed at dataset level and the same roles needs to be configured multiple times(5) for each dataset, because the datasets are unrelated (but have common relations of projectID) essentially it takes 25 times(5rolesx5datasets).
 
Task; is there a better way to store the data maybe in a single datset or in a dataflow so that the RLS does not need to be configured for each dataset each time? 
 
not sure if this is possible but ideally I want to be able to do rls for catogories of projects at a high level for all datasets and then rls for lower level at the dataset level such as all-view and restricted-view .
Need Help

1 Reply

  • ibarrau's avatar
    ibarrau
    Super User
    2 years ago

    Hi. Putting all tables together in a single content will depend on the permissions. The decision of creating a dataflow or dataset to keep the data might depend on the selfservice you want. First, consider if all developers should see the same tables from all the sources in order to know if you need only one component. Then think about how the report creators interact. If you have one dataflow, the creators will have more responsabilities because they will have the capacity of transform things with power query and the responsability of creating the measures and RLS for their models. They should have more seniority or expertise. On the other hand, if you have one dataset, all the ETL, measures and RLS is already created. They connect to a model that already has all the things they need to start creating visualizations.

    Of course it's not black or white. You could have the key RLS table at a Dataflow and import it in different models. You can have the first model for jira, the second model of excels and a third one like composite getting both models and the table of dataflow together. Then share with the correct audience.

    I hope that helps, there is no exact answer because it depends on the scenario