DATA WAREHOUSE SCHEMA EVOLUTION WITH EXTENDED HIERARCHY SEMANTICS
Sandipto Banerjee · OhioLink ETD Center (Ohio Library and Information Network) · 2007
A data warehouse technology provides a way to visualize data in a form that helps in the decision making process.This visualization is provided by a multi-dimensional schema that captures the user needs in terms of data content, constraints on data, and views of data.User requirements can change over a period of time and this causes the schema to be redesigned from scratch.Redesigning a schema is an expensive process in terms of resources and time.A solution to this problem is designing a schema evolution process that helps a schema to evolve.Models for conceptual design of data warehouse schemas have been proposed by xv 10.8 MDIH lattice 10.9 MDIH lattice for the Sub-lattice 10.10 Schema Supporting Additional Features 10.11 Multi-dimensional Lattice 10.12 Instances of Product Dimension 10.13 Sub-lattice 10.14 MDIH lattice 10.15 Instances of Product Dimension 10.16 Instances of Location Dimension 10.17 Non-covering hierarchy 10.18 MDIH Lattice 10.19 Instances of Product Dimension 10.20 Sub-Lattice 10.21 MDIH Lattice 10.22 Evolution Process 10.23 Example of an Evolution Process 10.24 Example Star Schema 10.25 Star Schema and Multi-dimensional Lattice 10.26 Evolved Star Schema and Multi-dimensional Lattice 10.27 Evolved Multi-dimensional Lattice 10.28 Evolved Multi-dimensional Lattice 10.29 Evolved Multi-dimensional Lattice 10.30 Multi-dimensional Lattice for a Star Schema 10.31 Evolved Multi-dimensional Lattice xvi 10.32 Evolved Multi-dimensional Lattice 199 Chapter 11 11.1 Evolution of Time Dimension 208 11.2 Product Dimension 210 11.3 Location Dimension 212 11.4 SalesCube 215 11.5 Original Schema 216 11.6 Evolved Schema 217 11.7 Test Dimension 221 11.8 Changes in Hierarchy When a Level is Deleted 222 11.9 Changes in Direct, Non-covering Hierarchy When a Level is Deleted 224 11.10 Changes in Direct, Non-covering, Non-onto Hierarchy When a Level is Deleted 226 11.11 Changes in Direct, Non-covering, Non-onto, Non-strict Hierarchy When a Level is Deleted 227 Chapter 12 xvii List of Tables Chapter 2 2.1 Additional Features of a Data Warehouse Chapter 3 region.By introducing a new level in the Warehouse dimension, an analyst can view the sales figures by regions and accordingly make decisions based on them.This example shows schema evolution as a result of adding a level to a dimension.In this research we design a multi-dimensional model that we call the generalized model along with evolution operators to support schema evolution to enhance decision making capabilities in a changing data warehouse environment. General Research ObjectiveOur research objective is to enhance decision making capabilities in a data warehouse environment by designing a generalized model to support schema evolution. Specific Research ObjectivesIn order to achieve the general research objective, we propose to answer the following specific questions: A. How are data warehouse schemas typically structured?What are the core features of a data warehouse data model?Are there any additional features of a data warehouse to support dimensional modeling?B. Is there a need for a formal model to describe a multi-dimensional model along with its core and additional features?Do contemporary data warehouse formal models support schema evolution?C. How can a generalized formal model be designed to support the core features of a data warehouse?How can the correctness of the model be verified?D. How is schema evolution supported over the core features of a data warehouse?E.How can the generalized model be implemented to illustrate schema evolution?How is schema evolution supported over the additional features of a data warehouse?Are there any new operators that help in this process?G. How do interactions take place between more than one dimension in a schema and how can aggregation be supported over the levels of the dimensions?H. How is aggregation supported at the instance level of a data warehouse?I. How is aggregation affected and supported in a schema evolving environment? Research MethodologyIn order to achieve the specific research objectives, we survey data warehouse research literature and conduct the following activities to design a data warehouse model that supports schema evolution and aggregation in an evolution enabled environment.A.Describe selected data warehouse models to illustrate dimensional modeling and identify the core features of a data warehouse.Survey literature to identify the