Research On Visible Multi-dimensional Data Modeling Technology Based On CWM

Data Warehouse as a new technology has been receiving more attention gradually, and brought great benefit to enterprises. On-Line Analytical Processing (OLAP) is the most important application for Data Warehouse, and it organizes data in Data Warehouse as a cube and supplies multidimensional analysis operations, so that decision-makers could analyse them dynamically and deeply.Multidimensional data model is the key technology of OLAP. By creating semantic objects of cube, like dimension, hierarchy, and measure etc, it satisfies the multidimensional modeling requirement of OLAP, and realizes the character of multidimensional analysis operations.This dissertation is focused on technologies of multidimensional data modeling and metadata management in OLAP domain, including: techologies of multidimensional data model, irregular dimension transformation, research on metadata management strategy, and designation of multidimensional modeling system etc. The main content and contribution in this dissertation include:1. By considering both the advantages and disadvantages of multidimensional data models presented today, this paper extended and improved the present multidimensional data model, and analysed the physical implementation of multidimensional model.2. In considering that traditional models could not support the irregular dimensions, this thesis analysed the present transforming algorithms for irregular dimensions, and presented two improved algorithms based on the shortages of the transforming algorithm for non-covering dimensions.3. We put forward an OLAP metadata management strategy based on XML technology, and designed a metadata adapter between CWM and our own system to exchange metadata information with chief DW manufacturers in market.4. Finally, we implemented a multi-dimensional modeling system on Java platform, which includes function modules of multi-dimensional data schema management, data source management and role management.