Research And Application On The Metadata In The Data Warehouse

Metadata is an important part in the data warehouse, metadata in the data warehouse exists in the process of creation, maintenance, management and usage of data warehouse. So the effective management of metadata in the data warehouse is one of the determinants on the success of the data warehouse project.This paper mainly analyses and summarizes the two existing metadata standards—the common warehouse model(CWM) and the open information model(OIM), and then after analyzing the situation and functions of metadata management, a principle of metadata management system design is proposed. Then this paper discusses the general architectures of metadata management, and after analyzing and summarizing the existing architectures, a mixed metadata management architecture is given.This paper researches on the prerequisite of data source integration in the data warehouse—data source mappings, and after the illustration of a mapping example, this paper summarizes the three activities of data source mapping. Finally, this paper proposes to use model management to manage the metadata in the data warehouse. This method is to use some high-level alberic operations to realize the effective management of metadata in the data warehouse, this paper complement the syntax and semantics of these operators, and proposes an algorithm of Match operator which is used for tree schemas—TreesMatch Algorithm and describes a typical Merge operation algorithm. In the end this paper uses examples of such typical problems of metadata management in the data warehouse as new data source integration and the creation of new schemas to prove its feasibility.