A Functional Data Model for Analytics

This paper presents a functional data model – a model that is fundamental to supporting management analytics applications. This functional data model, henceforth referred to simply as “the functional model”, is different from, but complementary to, the relational model. The functional model is also distinct from other similarly named concepts, including the DAPLEX functional database model, and functional language databases.

The functional model comprises multidimensional hierarchical consolidation, as is commonly found in OLAP technologies, relational-based and otherwise. But it goes beyond OLAP by requiring a spreadsheet-like cell orientation, and definition of cells calculated as functions of other cells. Because a functional definition of calculation according to formulas and consolidation is incorporated into the data model, databases implementing the model will inherently always return calculated cell values that are up to date and consistent with respect to the latest input cell values in interactive database update scenarios. These concepts are lacking in the relational model but are essential for support of flexible and interactive business performance management analytics.

The validity and effectiveness of the functional model is evident in the long-standing success of commercial product technology that embodies the model to deliver practical analytics solutions across a broad range of management analytics domains.

We present a formal definition of the functional model, a brief discussion of database technology that implements the model efficiently, and a comparison that shows how the relational model is inadequate in terms of expressiveness for scenarios fundamental to management analytics.