Research On The Multidimensional Data Model And Aggregation Algorithm In LE-OLAP

With the rapid development of database technology, people gradually realized that the information is so important to an enterprise, so companies begin to use a variety of database management systems to store the data of the company. In order to effectively use the data, people usually use OLTP (online transaction processing) technology to deal with these data. With the accumulation of data, enterprise decision-makers want to make better use of these historical data, and analyze the law to guide decision-making and company development began to use the data warehouse and OLAP (online analytical processing) technology to deal with the vast amounts of historical data to get useful information.OLAP can be divided into ROLAP and MOLAP according to the different organization of the data. ROLAP is based on relational database, and the data processing depends on the relational tables and SQL queries. And MOLAP is based on the multi-dimensional structure, data are usually organized in the form of data cubes, it can support the multidimensional view. At present most OLAP applications are ROLAP which is based on relational database. But compared with ROLAP, MOLAP is more flexibility to provide fast query, and allow users to observe the aggregation results from different angles more easily.In this paper, firstly, discuss the research status of multi-dimensional model and OLAP, and then select a feasible solution to design and implement the LE-OLAP multidimensional analysis engine. The research works are summarized as follows:1. Make a thorough study in the modeling of multidimensional model, and put forward the conception interested cube. Then establish a general multidimensional model with good scalability.2. By studying the common cube aggregation algorithm, determine the advantage and feasibility of the BUC aggregation algorithm in LE-OLAP. Then introduce the BUC algorithm, and improve the BUC algorithm by adding the levels and interesting cube to aggregate calculating, and realize the multi-thread BUC and block BUC.3. Apply the improved BUC algorithm to the LE-OLAP, analyze and compare the experimental results.