Since the public preview announcement in June ‘19, organizations are increasingly relying on shared and certified datasets to help their users build reports based on trustworthy and authoritative data. Now, we’re excited to announce general availability (GA). We’ve also added several improvements based on customer feedback.
With shared and certified datasets, Power BI users can easily discover and reuse enterprise-wide semantic models. They can be shared across workspaces, fostering reuse and maintaining the single source of truth. An updated permission model ensures that only the right users get access to those datasets. They can also be promoted and certified as trusted and authoritative datasets. As the following screenshot illustrates, promoted and certified datasets show up prominently in the dataset discovery experience to steer report authors effortlessly toward these high-quality sources of data. Dataset owners can promote their datasets when they are ready for further exploration by others. The ability to certify datasets, on the other hand, can be tightly controlled and documented internally via a new tenant setting in the Power BI admin portal. For more information, see Promote your dataset and Certify datasets in the product documentation.
As a dataset owner, you can perform impact analysis to identify those reports that consume your shared datasets across your organization. This is particularly useful if you plan to modify or delete a shared dataset. For example, it might be a good idea to rebind any downstream dependencies to an alternative dataset before performing a delete operation. For more information about impact analysis, see Dataset impact analysis in the product documentation.
During the public preview, Power BI did not allow a dataset owner to delete a shared dataset with existing downstream dependencies in other workspaces. This restriction has now been lifted. As you can see in the following screenshot, Power BI warns you that all reports and other artifacts depending on this dataset within the same workspace will also be deleted, and then informs you that there are also reports in other workspaces. You then have the option to perform impact analysis or proceed with the delete operation.
It is important to point out that improved capabilities for dataset impact analysis have just been announced, By using the lineage view, you can easily determine the upstream and downstream dependencies and take appropriate action.
If you decide to delete a shared dataset, any dependent reports in other workspaces will no longer work because the underlying dataset no longer exists. Users will see an error indicator with a tooltip stating that there is no dataset associated with the report, as in the below screenshot. The user must rebind the report to an alternative dataset or delete the report if no alternative dataset is available.
Because broken reports impact the user experience, it is not recommended to delete shared datasets without impact analysis. It is better to work with report authors to rebind their reports first and then delete the shared dataset in question in a subsequent step. Only delete shared datasets with existing downstream dependencies if the situation calls for immediate action.
It is important to remind you that shared and certified datasets are only available with the new Workspace experience, and not the classic Workspaces based on Office 365 Groups, so make sure you upgrade your workspaces to the modern experience.