【摘要】 随着大数据时代的到来,数据查询迎来了巨大的挑战。首先,单个数据查询任务应具有更快的处理速度,以应对数据量的爆炸式增长;其次,数据查询的复杂性增加,导致DBA人工调整方法以及自动配置工具都无法保证配置的有效性和及时性。传统数据查询过程中,不考虑数据的潜在分布特征。机器学习技术可构建数据依赖型模型,使得数据查询的速度和空间利用率均优于传统方法。本文对将机器学习技术使用于数据查询优化的研究工作进行总结和归纳,主要包括索引结构优化、查询过程优化和学习型数据库系统调优。在对已有技术分析基础上,指出机器学习在数据查询的未来研究方向和可能面临的挑战。更多还原

【Abstract】 With the advent of the big data era, data querying poses significant challenges. First, a single data query task should have faster processing speed to cope with the explosive growth of data volume; second, the increased complexity of data queries leads to the inability of DBA manual tuning methods as well as automatic configuration tools to ensure the validity and timeliness of configuration. The traditional data query process does not take into account the underlying distribution characteristics of the data. Machine learning techniques allow the construction of data-dependent models that enable data queries to be faster and more spatially efficient than traditional methods. This paper summarizes and synthesizes research work on the use of machine learning techniques for data query optimization, including index structure optimization, query process optimization, and learning database system tuning. Based on the analysis of existing techniques, the future research directions and possible challenges of machine learning in data querying are pointed out.更多还原