Relational Blending

The dynamic blending of data stored in a relational database with data stored/derived by the in-memory analytic engine is supported. 

Relational blending is a data management approach.  This enables analytic modelers to better manage the trade-offs between building unsustainable/poor performing complex analytic models that contain too much detail and the need to report on that level detail (data that may be transactional in nature or constantly changing).  Relational Blending makes it possible to seamlessly integrate detail data points into the analytic reporting process.

Relational Blending Types

There are three methods of relational blending.

1. Drill-Back Blending (One-to-Many Relationship)
   This method of relational blending is used to provide access to detailed information that does not exist in the analytic model.  This capability is delivered right out-of-the box with its predefined drill back to stage detail data.  In addition, the stage integration engine provides drill-back and drill-around capabilities against external data.

2. Application Blending (One-to-Many Relationship)
   This method of relational blending leverages the OneStream MarketPlace Specialty Planning and Compliance applications.  This collects information in a transactional register format and seamlessly maps/loads summarized data into an analytic model.  These applications also provide predefined transactional level reports as well as predefined drill-back connectors allowing drill-down from a summarized analytic model to the detailed register transaction data.

3. Model Blending (One-to-One Relationship)
   This method of relational blending combines the power of the in-memory analytic engine with the flexibility of relational database storage.  This functionality is provided as part of the Finance Engine API and the functions can be found under the api.Functions path in a Finance Business Rule.