Research And Implementation Of Query System Based On Materialized Views

It is very important to quickly answer queries in database environment. Besides query optimization, another efficient way is query rewriting using materialized views. The definitions and results of materialized views are both stored in database. Users give their queries against base tables, the queries are transparently rewritten using materialized views and the time consumed by queries is reduced by avoiding accessing huge raw records and performing time-consuming operations.As far as application systems with minitype database are concerned, minitype database could not provide the mechanism of materialized views and query rewriting. When queries contain join and aggregate operations with large amount of data, time consumption problem as a bottleneck could not be solved. Therefore, the paper studies Query System Based on Materialized Views considering minitype database. Queries are rewritten using materialized views via the Query System and then performed efficiently in minitype database.First, the paper studies conditions and algorithms of query rewriting, and using the property of join relation with foreign key extends the usability of materialized views; the idea of classification is used to derive the WHERE conditions of queries, and how to derive the query conditions and how to generate the compensated conditions are specified by the materialized view conditions. The required conditions and the whole algorithm of query rewriting based on foreign key relation are proposed.Second, applying Object-Oriented method, Query System Based on Materialized Views in B/S structure is implemented on Windows OS using Java Language, and its core is the query rewriting algorithm proposed above. Several query instances are tested on the Query System successfully, which proved the correctness of query rewriting algorithm based on foreign key.Finally, Query System Based on Materialized View is applied on KnowIT which is an analysis and show system of passengers flow. The core program of Query System is transplanted to KnowIT, and the results of efficiency comparison experiment are analyzed between queries on base tables and queries on materialized views, which proved the validity of Query System Based on Materialized View.