Workspace datasets

As per the staged deployment lifecycle diagram, this architecture requires distinct Power BI datasets per app workspace. To minimize resource usage and for data security reasons, the development workspace dataset could include the minimal amount of data necessary and exclude all sensitive data. This would allow the organization to comfortably provide development access to teams of content developers, potentially from outside of the organization. Access to the test workspace could be limited to a small number of trusted or approved users within the organization and thus could include sensitive data. Finally, the production workspace dataset would have the same schema as the other datasets but include the full volume of data as well as sensitive data.

If a common schema exists between the different datasets in each workspace, the source dataset of a Power BI Desktop report file can be revised to a dataset in a separate workspace as per the Switching source datasets section in Chapter 11, Creating and Formatting Power BI Reports.

For example, the report file (.pbix) approved for migration from the development workspace to the test workspace could be opened, modified to reference the test workspace dataset, and then published to the test workspace. This approach represents a manual alternative to the Power BI REST API described in the following section.

A new feature is expected in 2018 that will allow a Power BI report to reference a dataset in an external app workspace. The availability of this feature will help eliminate the resource cost and manageability issues of duplicated datasets across multiple app workspaces.

For example, distinct Power BI apps developed for the finance, sales, and marketing teams could all leverage a single production dataset in a dedicated workspace rather than individual datasets within each workspace. The availability and implementation of this feature will revise the architecture of staged deployments of Power BI content via large Power BI datasets.

Another alternative to avoid the duplication of a dataset across multiple apps is Analysis Services. With Analysis Services, either on-premises via SSAS or in the cloud via AAS, Power BI reports can be created with Live connections to development, test, and production data models. Information on utilizing Analysis Services and its advantages as the data modeling tool and engine for Power BI is included in Chapter 19, Scaling with Premium and Analysis Services.