首页 | 本学科首页   官方微博 | 高级检索  
     检索      

时序数据多维聚合查询服务的实现
引用本文:盛家,房俊,郭晓乾,王承栋.时序数据多维聚合查询服务的实现[J].重庆大学学报(自然科学版),2020,43(7):121-128.
作者姓名:盛家  房俊  郭晓乾  王承栋
作者单位:北方工业大学 大规模流数据集成与分析技术北京市重点实验室, 北京 100144;北方工业大学 数据工程研究院, 北京 100144
基金项目:国家自然科学基金面上资助项目(61672042)。
摘    要:随着电能质量监测点不断扩大,产生海量具有时序特性的多维电能质量数据,当前的诸多数据查询方法不能适应电网电能质量监测数据的交互式多维聚合查询需求。研究提出时序数据多维聚合服务的实现方法,为内存中预聚合后的任务结果建立哈希存储结构,对实时数据建立位图索引存储结构,将历史数据的预聚合数据尽量存储于内存中,改进随机读写的低性能问题,提升查询效率,解决交互式查询问题。同时运用最优聚合任务算法选择出尽量多的预聚合任务数,提高交互式查询命中率。实验验证了该算法的可行性,与分组二维背包算法相比,在预聚合任务数量选择方面具有一定优势。

关 键 词:时序数据  聚合查询  预聚合  交互式查询
收稿时间:2020/1/20 0:00:00

Implementation of multidimensional aggregate query service for time series data
SHENG Ji,FANG Jun,GUO Xiaoqian,WANG Chengdong.Implementation of multidimensional aggregate query service for time series data[J].Journal of Chongqing University(Natural Science Edition),2020,43(7):121-128.
Authors:SHENG Ji  FANG Jun  GUO Xiaoqian  WANG Chengdong
Institution:Beijing Key Laboratory on Integration and Analysis of Large-scale Stream Data, North China University of Technology, Beijing 100144, P. R. China;Institute of Data Engineering, North China University of Technology, Beijing 100144, P. R. China
Abstract:With the continuous expansion of power quality monitoring points, a large number of multi-dimensional power quality data with time series characteristics have been generated. The existing data query methods can not meet the need of interactive multi-dimensional aggregation query of power quality monitoring data. This paper presents a method to implement multi-dimensional aggregation service for sequential data. It establishes a hash storage structure for pre-aggregated task results in memory, a bitmap index storage structure for real-time data, and stores pre-aggregated historical data in memory as much as possible thereby improving the performance of random reading and writing, and the efficiency of query, solving the problem of interactive query. At the same time, the optimal aggregation task selection algorithm is used to select as many pre-aggregation tasks as possible to improve the hit rate of interactive queries. Experiments verify the feasibility of the proposed algorithm. Compared with the grouped two-dimensional knapsack algorithm, it has certain advantages in the number of pre-aggregated tasks.
Keywords:time series data  aggregate query  pre-aggregation  interactive query
本文献已被 CNKI 等数据库收录!
点击此处可从《重庆大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《重庆大学学报(自然科学版)》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号