| Summary: | 碩士 === 大同大學 === 資訊工程學系(所) === 92 === Data cube materialization is commonly used in reducing OLAP response time. However, to materialize a whole data cube needs large disk space, because the focus on the interested subjects of decision makers results in only a small portion of data cubes being frequently accessed. The subcube, a finer partition of a data cube is proposed. The subcubes are formed by mapping the multi-dimensional queries, and the number of subcubes becomes larger after various queries issued by users with different dimension levels and value ranges. The management framework of these subcubes is important. Another concerned issue is how to pick up the most appropriate subcubes for the query processing. The technique directly affects the query performance. This thesis research uses binary trees and linked lists to manage and organize the subcubes. For saving the query processing time, an algorithm for searching appropriate subcubes is proposed, and the performance of the subcube based query processing algorithm is improved.
|
|---|
