数据仓库中雪花模式的Skyline-Join查询

Skyline查询能够有效地实现多目标最优化,而数据仓库中的OLAP也是针对多维数据进行分析,因此,针对Skyline查询在数据仓库中的应用,提出了数据仓库中雪花模式的Skyline-Join查询算法.该算法首先将子维表M-Join父维表,然后渐进选择式地对事实表和父维表进行连接.每次连接之前都对事实表进行分组和组内Skyline计算,删除组内非Skyline元组,这样可以减少许多不必要的连接操作,使得查询效率大大提高.通过实验证明,在事实表元组数量逐渐变大和维表个数逐渐增多的情况下,提出的算法比先Join后Skyline计算的naive算法效率上有明显改善.

Skyline-Join in Snow Schema of Data Warehouse

Skyline query is efficient in multi-criteria decision making systems,while OLAP in data warehouse mainly focuses on the analysis of multi-dimension data.Therefore,a progressive and selective skyline-join algorithm is proposed to apply skyline query into snow schema of data warehouse.This algorithm first M-Joins child dimension tables with parent dimension tables,then progressively and selectively joins fact table with each parent dimension table.Before each join,fact table will be grouped and skyline computation will start in each group;and then the non-skyline tuples in each group will be pruned to avoid many unnecessary join operations.That can greatly improve the algorithm's efficiency.The experiments demonstrate that when the size of fact table and the number of dimension tables increases,the proposed algorithm is much more efficient than the naive skylinejoin algorithm.