Compositional Dependencies and High-Order Relations Formulas at Relational Data Models

Abstract

It is quite difficult to abandon well-ordered formal constructions, even if they are of little use in practice. The normalization method which, with the emergence and development of an alternative semantic direction, has gradually faded into the background has today found itself in such a situation in data modeling. Nevertheless, the practical potential of normalization has not been exhausted. Removing the restriction on the mandatory decomposition of factographic relations (FRs) with join dependencies and expanding the range of compositional dependencies that can objectively be present in such relations significantly expand the boundaries of the method’s suitability, including allowing one to fill in the gaps in the incompleteness of information concerning the organization and functioning of the domain being modeled. This paper focuses on identifying and considering new classes of dependencies: projection, Cartesian product, intersection, union, and difference, as well as on deriving a formula for an arbitrary FR that comprehensively takes into account the influence of all compositional dependencies found in the relation on its possible content.