首页> 外文会议>SIGMOD/PODS >Optimal Top-Down Join Enumeration
【24h】

Optimal Top-Down Join Enumeration

机译:最佳自上而下的加入枚举

获取原文
获取外文期刊封面目录资料

摘要

Most contemporary database systems perform cost-based join enumeration using some variant of System-R's bottom- up dynamic programming method. The notable exceptions are systems based on the top-down transformational search of Volcano/Cascades. As recent work has demonstrated, bottom-up dynamic programming can attain optimality with respect to the shape of the join graph; no comparable re- sults have been published for transformational search. How- ever, transformational systems leverage benefits of top-down search not available to bottom-up methods. In this paper we describe a top-down join enumeration algorithm that is optimal with respect to the join graph. We present performance results demonstrating that a com- bination of optimal enumeration with search strategies such as branch-and-bound yields an algorithm significantly faster than those previously described in the literature. Although our algorithm enumerates the search space top-down, it does not rely on transformations and thus retains much of the architecture of traditional dynamic programming. As such, this work provides a migration path for existing bottom-up optimizers to exploit top-down search without drastically changing to the transformational paradigm.
机译:大多数当代数据库系统使用System-R自下而上的动态编程方法的某些变体执行成本的加入枚举。值得注意的例外是基于Volcano / Cascades的自上而下变换搜索的系统。正如最近的工作所示,自下而上的动态编程可以相对于连接图的形状达到最佳状态;没有可比较的可比素,用于转型搜索。然而,变革系统利用自上而下的搜索的益处,无法用于自下而上的方法。在本文中,我们描述了一个自上而下的连接枚举算法,其对连接图是最佳的。我们呈现绩效结果表明,使用分支和绑定等搜索策略的最佳枚举和算法产生的绩效表现得比先前在文献中描述的算法明显快。虽然我们的算法以自上而下的枚举搜索空间,但它不依赖于转换,因此保留了传统动态编程的大量体系结构。因此,此工作为现有的自下而上优化器提供了迁移路径,以利用自上而下的搜索,而不会对变换范例进行大幅更改。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号