Conceptual level design of object oriented data warehouse: Graph semantic based model

Abstract

This paper proposes a Graph semantic based Object Oriented Multidimensional Data Model (GOOMD). The model defines a set of graph based formal constructs that are used to specify the conceptual level design of data warehouses (DW). Besides offering a novel graph based semantic for multidimensional data representation, the proposed GOOMD model renders a unified realization of different logical structures of DW like star, snowflake, constellation, etc. The proposed simple but powerful algebra for GOOMD provides an efficient operational model for Online Analytical Processing (OLAP) to realize different OLAP operations like roll-up, drill-down, slicing, dicing, drill-across and drill-through. The proposed GOOMD model also supports a formal representation of advanced concepts like parallel hierarchies, many-to-many associations, additivity constraints etc.