针对基于delta表存储的what-if查询处理实现技术中查询性能较低的问题,对what-if分析的两个阶段提出优化方案:在假设视图生成阶段,提出基于Bloom filter算法实现what-if视图生成;在基于假设视图的OLAP查询阶段,提出基于Redis的查询缓存算法WFQC,提高基于假设视图的OLAP查询效率。实验结果表明,Bloom filter算法生成的what-if视图比使用传统的表连接算法生成的what-if视图性能提高了约32%;与传统的物化视图实现机制相比,基于Redis查询缓存的OLAP查询性能提升了2倍。
Since the performance of what-if query processing often tends to be time consuming,a two phase optimization schema for what-if analysis was put forward.In the phase of what-if view generalization,a Bloom filter based algorithm was proposed to generate what-if view.In the what-if view based OLAP query processing,the query caching technique WFQC to improve the effi-ciency of what-if analysis was presented.Experimental results show that the performance of what-if view generalization is im-proved by 32% than traditional join algorithms and the query performance of the proposed methodology is twice as materialized views based approach.