Since the queries are dynamic, with unknown selection criteria at compile time, plan generation can be incorrect, even with the availability of full distribution statistics on the data.
由于查询是动态的,在编译时并不知道其选择标准,因此,即使有了关于数据的完整的分布统计,仍然可能生成错误的计划。
Figure 5 shows the retrieval of candidate service statistics and the dynamic selection and late binding to a service endpoint.
图5展示了对备选服务统计数据的检索以及某个服务端点的动态选择和延迟绑定。
Dynamic SQL statements automatically use these updated statistics.
动态sql语句自动使用这些更新的统计数据。
应用推荐