Functional Dependency Trees for BI Abstraction Layers

The patent introduces functional dependency trees as an intermediary structure between the abstraction layer and the physical database schema. This mediator captures relationships between dimension objects (such as transitive dependencies and fan traps) that are lost in the orthogonal abstraction layer, while still allowing users to query using simplified analysis objects. The functional dependency tree acts as a bridge that preserves relationship information without exposing database complexity to users.

The patent implements nested functional dependency trees where dimension objects can have dependencies on other dimension objects through multiple levels of analysis objects. The functional dependency tree structure allows dimension objects from different analysis objects to be linked through transitive dependencies, creating a nested hierarchy that represents complex relationships while maintaining manageable complexity through the tree structure.

The patent extracts relationship information from the physical database schema and represents it separately in functional dependency trees within the abstraction layer. By taking out the relationship representation from the complex static alias approach and placing it in a dedicated functional dependency structure, the system maintains explicit relationship representation while reducing overall abstraction layer complexity. Users can navigate relationships through the functional dependency tree without dealing with complex alias management.