Datamart is the new feature announced at Build. However, Datamart is not solution for everything. It is important to understand what Datamart, Dataflows or Datasets are and when which tool should be used or can be used.
Dataset
A dataset is the main entity in Power BI. They represent data sources that is ready for use in report creation and visualization. Datasets are created when creating push datasets, streaming datasets, when uploading Excel workbooks or directly with the report (*.PBIX file).
The following information is stored in a dataset:
- The connection to the data source
- Tables and their data
- Calculated columns, tables and measures
- Hierarchies
- Formatting and settings of fields
- Relations
Datasets are available in all versions of Power BI, including the Free version.
Dataflow
A Dataflow is a collection of tables created and managed in the Power BI service. They are a kind of self-service data preparation for Big Data in Power BI. After the Dataflow is created, you can use Power BI Desktop or Power BI Service to create datasets, reports, dashboards and apps.
Dataflow is often used in reusable data preparation (ETL) that is used in multiple datasets or reports. This can create a singe source of truth by requiring data analysts to connect to the Dataflow rather than the underlying systems.
Dataflow is suitable for tables that are used frequently (calendars, personnel data, hierarchies, groupings, etc.) or when using Azure Data Lake (Dataflow was made for this 😉).
Dataflow is available from the Power BI Pro version.
Datamart (Preview)
Power BI Datamart is a self-service database which is strictly speaking an Azure SQL database that can be created by the Power BI user without a database administrator. Data can be loaded from various data sources, extracted using Power Query, then transformed and then saved. Aggregate functions, grouping, etc. can be created either with Power Query or SQL.
Datamarts are used by users who do not have access to Power BI Desktop, want to make data available quickly via an SQL endpoint, or want to perform ad-hoc analyses on the web, regardless of the data source.
Datamart is included in the Power BI Premium Capability and Premium Single User License version.