Concurrent Data Materialization for Object-Relational Database with Semantic Metadata

Abstract

For a company with many databases in different data models, it is necessary to consolidate them into one data model interchangeable and present data in one data model concurrently to users. The benefit is to let user stick to his/her own data model to access database in another data model. This paper presents a semantic metadata to preserve database constraints for data materialization to support user's view of database on an ad hoc base. The semantic metadata can store the captured semantics of a relational or an object-oriented database into classes and stored procedures triggered by events. The stored constraints and data can be materialized into a target database upon user request. The user is allowed to perform the data materialization many times alternatively. The process can provide a relational as well as an object oriented view to the users simultaneously. This concurrent data materialization function can be applied to data warehouse to consolidate heterogeneous database into a fact table in a data model of user's choice. Furthermore, a user can obtain either a relational view or an object-oriented view of the same dataset of an object-relational database interchangeably.