Research And Application On Grouping Aggregation Algorithm For OLAP Based On Dimension Hierarchical Encoding

With the development of the heat charging system reform and thermal charges systeminformationization, running transactional database stored a large number of detailed,valuable historical data, how to analyze these massive data deeply and rapidly extractinformation useful for thermal business decision analysis, becomes the main challengesfacing the heat charging management system decision-makers. To solve the problem, agrouping-aggregation algorithm for OLAP based on dimension hierarchical encoding isproposed in this paper, and applied to the heat charging management system.Firstly, the research status for data warehousing and OLAP is summarized in thispaper based on lots of related literatures. The basic theoretical knowledge of datawarehouse and OLAP is introduced. And typical OLAP algorithms are researched.Secondly, to adapt to the needs of the current heat charge reform in China, a thermalcharging management system is developed, which increases the heat metering modules onthe basis of original area of charging management modules. So the thermal chargingmanagement system achieves the target that area charges and heat metering chargesunified management. And according to the characteristics of data mass in the systemdatabase, a thermal toll collection management system data warehouse is designed andimplemented in this paper.Thirdly, since the thermal charging management system data warehouse has largevolumes of data, and its’ query conditions are complex and its’ dimensions have differentHierarchy. To improve its query efficiency, a grouping-aggregation algorithm for OLAPbased on dimension hierarchical encoding is proposed in this paper. The algorithm uses abinary encoding to compress data. First, according to the query condition, the dimensionhierarchical grouping attribute set of each dimension table is got, and the dimensionhierarchical grouping attributes encodings of different levels in each dimension table areobtained by finding the corresponding hierarchical tree. Then the dimension hierarchicalgrouping attribute set encodings are calculated and put into temporary table to sort andgroup. Last, the grouping sequence numbers of records in fact table are calculated, and the buffer was located to finish grouping aggregation calculation. The analysis is made fromthe CPU time and I/O times.Finally, the charges situation is analyzed through OLAP queries, and the thermalcharge system data warehouse is as a data source to experiment. The experimental resultsverify the validity of the GABDHE algorithm.