Research Of Query Optimization And Index Of Data Warehouse

Business intelligence market scale flourishes unconventionally. The three main supporting technologies of BI, data warehousing, online analytical processing and data mining, are also the hot topic of academic research and discussion. Data warehouses which stores volumes of historical data, are the structured data environment of decision support system and online analytical processing, and online analytical processing also involves the complex queries in mass data. Therefore, to enhance the efficiency of data warehouse queries becomes one of the key issues of improving the efficiency of data warehousing applications and online analytical processing applications.This dissertation introduces the theoretical basis of data warehouse and online analytical processing technology at first, then the discussions focus on the analysis of data warehouse's query performance. There are mainly three methods to solve the problem at current: materialized view, query optimization and index technology. This paper introduces the classification of index and gives a detailed description on basic concepts, principles and methods of commonly used index technologies in data warehouse, including the B tree, B- tree, B+ tree, B*tree, bitmap, inverted index, join index and hash table.Based on the above, this paper proposes a new integrative index model which is called BBI and gives the methods to build and use the model. Based on the feature of Data Warehouse and OLAP queries, this index model is built with inverted index, aggregation table, bitmap index and b-tree. It not only greatly promotes the efficiency of frequently used queries but also improves the performance of other queries. At last a further optimization method is proposed to improve the index model and analyses are given on the time and space performance. The analytical and experimental results demonstrate the utility of BBI.