Research On Materialized Views Self-maintenance Based On Data Warehouse

With the fast development of computer technologies and enterprise increasing needs, data warehousing technologies come out, which are based on a comprehensive and prefect information management system. A data warehouse collects and maintains a large of data from several distributed and heterogeneous data sources, and often the data is stored in the form of materialized views in order to provide fast access to the integrated data. However, when the sources change, the views in the warehouse must be consistent with the base data, and the views must be maintained.The query responding time can be speeded by pre-storing. The performance of the data warehouse has been improved by using and studying materialized views. Although great achievements of technologies of materialized views maintenance research have been made, there are many problems of materialized views maintenance technology because of the consistency of materialized views characteristics. This paper focuses on the materialized views maintenance, sums up the method materialized views, the advantage of materialized views, the maintenance processes of materialized views, analyzes the strategies of materialized views in detail. In the dissertation there adopts incremental maintenance techniques, combined with the self-maintenance method according to auxiliary views, proposes an algorithm of materialized views self-maintenance, and with the algorithm the example detection efficiency is made. Their results show it is high efficient and can largely save lots of space of materialized views.Considering not all the views can be self-maintained, combined with the detected algorithm of judging whether views can be self-maintained or not, this paper proposes a detected UD algorithm in relation with views update. This UD algorithm can avoid network transmissions and make sure the accordance between materialized views and base data sources according to visiting auxiliary views, which can help to improve the reply time and system performance.Because it needs a large of space to store materialized views, this paper proposes a computable model and show its optimized and updated process, based on reducing the space of materialized views. This model can physically delete the useless views, according to the strategy module of the model.