Are Hybrid Tables tied to a developer license type?
No. Incremental Refresh and Hybrid tables are tied to workspace (dataset) type. They are set up in Power BI Desktop. A developer must have Pro or Premium Per User (PPU) license to publish the report to the service. See the next two paragraphs for workspace (dataset) limitations.
Are Hybrid Tables available for Non-Premium datasets/workspaces?
No. While it is possible to set up Incremental Refresh for datasets published to a non-premium (shared) capacity, Hybrid tables are not supported in this scenario. Power BI Desktop gives a warning message when you try to do that:
“At least one of your tables has a combination of incremental refresh and real-time data, which can only be published to Premium workspaces.”
It is interesting that publishing a report with Incremental refresh and hybrid options (a direct query partition) to a non-premium workspace does not generate any errors; however, since XMLA endpoint which allows tools such as SSMS or Tabular Editor to connect to a dataset is not supported in a non-premium workspace, there is no way to look at the partitions created even for incremental refresh.
Are Hybrid Tables available for Premium Per User (PPU) /Power BI Embedded datasets/workspaces?
Yes. Both Incremental Refresh and Hybrid tables are available for Premium Per User (PPU) and Power BI Embedded datasets/workspaces.
Note: Only users who have a Premium Per User (PPU) license can publish to or access reports from a PPU workspace.
Can the Direct Query partitions of a Hybrid table have a different data source than the rest of partitions?
All partitions must query data from a single source. When setting up Hybrid Tables from Power BI Desktop, all partitions will use the same underlying table by default. For more advanced scenarios where Tabular Model Scripting Language (TMSL) or Tabular Object Model (TOM) are used to set up custom partitions, all partitions must still come from a single data source. The question is: is the data source considered a database or a single table? Can each partition come from a different table? I have not tested this scenario yet.
Are Hybrid Tables supported in On-Prem Power BI Report Server?
No. Even composite models are not supported in the On-prem Power BI Report Server. Comparing Power BI Report Server and the Power BI service – Power BI | Microsoft Docs Additionally, the Automatic page refresh for DirectQuery models is not allowed. While not required, this feature is an essential part for the scenario where more recent data from a DirectQuery partition shows updates in a report without interaction for the user using the automatic page refresh.
Are Hybrid Tables supported in SQL Server Analysis Services (SSAS)?
No. There is no option in the interface that lets you define incremental refresh with or without a Direct Query partition; however, in SSAS you can define partitions and manage them yourself.
Are Hybrid Tables supported in Azure Analysis Services (AAS)?
No. There is no option in the interface that lets you define incremental refresh with or without a Direct Query partition; however, for Azure Analysis Services you can define partitions and manage them yourself. using tools such as Tabular Editor. Here is blog post on this topic.
Are Hybrid Tables supported in Data Flows?
Incremental refresh is supported in Data Flows. Hyrbid tables unknown.
Is it possible to set up Incremental Refresh for more than one table?
Yes. Even though all tables must use the same RangeStart and RangeEnd parameters, each table can have its own partitioning strategy. For example, one table can be partitioned by Year and another one by Quarter. The question is: is it possible to define different ranges if you use Tabular Editor or another tool.
Can a Hybrid table be set up in Tabular Editor 3 (paid version)?
Patrick Lablanc (Guy in a Cube) has a video here that shows you how to set up a hybrid table using Tabular Editor 3. This example shows you how to set up a hybrid table that is the opposite of what you can do with the built in GUI from Power BI Desktop. In this example, the historical data is in DirectQuery and the new data is in Import.