| PMC: Select Materialized Cells in Data Cubes |
| |
| 作者姓名： | Hong-Song Li and Hou-Kuan Huang |
| |
| 作者单位： | School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, P.R. China |
| |
| 基金项目： | Supported by the National Key Scientific and Technological Project： Research on the Management of the Railroad Fundamental Information （Grant No.2002BA407B01-2） and the Science Foundation of Beijing Jiaotong University （Grant No.2003SZ003）. |
| |
| 摘 要： | QC-Tree is one of the most storage-efficient structures for data cubes in an MOLAP system. Although QC-Tree can achieve a high compression ratio, it is still a fully materialized data cube. In this paper, an improved structure PMC is presented allowing us to materialize only a part of the cells in a QC-Tree to save more storage space. There is a notable difference between our partially materialization algorithm and traditional materialized views selection algorithms. In a traditional algorithm, when a view is selected, all the cells in this view are to be materialized. Otherwise, if a view is not selected, all the cells in this view will not be materialized. This strategy results in the unstable query performance. The presented algorithm, however, selects and materializes data in cell level, and, along with further reduced space and update cost, it can ensure a stable query performance. A series of experiments are conducted on both synthetic and real data sets. The results show that PMC can further reduce storage space occupied by the data cube, and can shorten the time to update the cube.
|
| 关 键 词： | 数据库 多维数据库 数据挖掘 存储结构 MOLAP结构 |
| 收稿时间： | 2005-05-09 |
| 修稿时间： | 2005-05-092005-11-30 |
PMC: Select Materialized Cells in Data Cubes |
| |
| Hong-Song Li and Hou-Kuan Huang.PMC: Select Materialized Cells in Data Cubes[J].Journal of Computer Science and Technology,2006,21(2):297-305. |
| |
| Authors: | Hong-Song Li Hou-Kuan Huang |
| |
| Affiliation: | (1) School of Computer and Information Technology, Beijing Jiaotong University, Beijing, 100044, P.R. China |
| |
| Abstract: | QC-Tree is one of the most storage-efficient structures for data cubes in an MOLAP system. Although QC- Tree can achieve a high compression ratio, it is still a fully materialized data cube. In this paper, an improved structure PMC is presented allowing us to materialize only a part of the cells in a QC-Tree to save more storage space. There is a notable difference between our partially materialization algorithm and traditional materialized views selection algorithms. In a traditional algorithm, when a view is selected, all the cells in this view are to be materialized. Otherwise, if a view is not selected, all the cells in this view will not be materialized. This strategy results in the unstable query performance. The presented algorithm, however, selects and materializes data in cell level, and, along with further reduced space and update cost, it can ensure a stable query performance. A series of experiments are conducted on both synthetic and real data sets. The results show that PMC can further reduce storage space occupied by the data cube, and can shorten the time to update the cube. |
| |
| Keywords: | data warehousing data cube multidimensional database partial materialization |
| 本文献已被 CNKI 维普 万方数据 SpringerLink 等数据库收录！ |
| 点击此处可从《计算机科学技术学报》浏览原始摘要信息 |
|
点击此处可从《计算机科学技术学报》下载全文 |
