多路方体聚集完全立方体计算算法

摘要

The pre-computation of data cube is crucial to improving the efficiency of OLAP performance. Based on using the multi-way array aggregation full cube calculation algorithm as the reference, this article presents an algorithm of full cube calculation driven by the data result set. While a cube is completed the scanning, the algorithm implements the calculation of aggregation values of the new cubes which are generated by rolling up multiple dimension levels, therefore the multi-way cube aggregation is achieved. This algorithm supports calculating the cube with large data scale. The application results show that this algorithm is feasible and easy to implement.%数据立方体的预计算对于提高联机分析处理性能至关重要.在借鉴多路数组聚集完全立方体计算算法的基础上,提出了利用数据结果集驱动的完全立方体计算算法.算法在扫描完成一个方体的同时,完成方体沿各个维攀升形成的多个新方体的聚集值的计算,从而完成多路方体聚集.该算法支持大数据量立方体的计算.应用结果表明,算法可行,且易于实现.