Via the OneStream XF Workflow interface, the BI Blend solution can be opened. A read-only aggregate storage model will be showed, and this model justifies the source data for uniform and standardized reporting. The BI Blend data is stored in a new relational column store table, which is outside of the OneStream XF Database.

The following topics popped up during our tests and research:

• Fast aggregation for (very) large amounts of data
  • No eliminations are performed
• Single reporting currency translation
  • Direct translation only
• Per BI Blend process 1 reporting currency
• Leveraging Reporting, Integration and OneStream XF Metadata tools.

• • These tools don’t utilize relational properties, like percentage consolidation and aggregation weight.
  • Via Derivative Transformation rules additional aggregation points can be included.
• Executed outside the cube.
• Not support for view dimension.
• Basic time math (via helper rules).
• Multi-period data can simultaneously be loaded, up to 12 periods.
  • BI Blend is a CPU and Memory intensive process.

To integrate transactional data, the Stage engine is used. The BI Blend source data will contain additional members and values that are defined in the integration as attribute members.

BI Blend will utilize the Cube dimensions to transform the records into a unified reporting format (including aggregated values), which is stored in a structured relational table for OLAP reporting. This table is available via the standard OneStream BI Dashboard and Grid reporting tools as well as relational blending into Cube views and dashboards. The output are read-only fact tables optimized for reporting with high level compression. These tables are created by the BI Blend process automatically.

The Dimensions the BI Blend Engine are defined by the cube’s integration settings. Defining the Blend Unit dimension is a key decision. Thereby unique top members are preferred across these dimensions. BI Blend supports Extensible Dimension hierarchies, tested and designed up to two levels of nested cubes. The workflow import channel shows the BI Blend settings.

A learning mode will automatically run to optimize the processing when the initial processing is running. You should enable this again if the Blend Unit has been changed.

We’ve noticed the following:

• Data access and other data cell level controls are not supported in BI Blend.
• The validation is limited to target mapping. Data intersection validation is not performed.
• A log file is created and intended to provide statistics to manage the blend task.
• Only UD8 can be used to generate Attribute Dimension members for aggregation.