sql server执行计划 - 哪个查询性能很好

Question

我正在学习SQL Server 2008执行计划并进行测试我已经运行了两个查询并尝试比较结果。

第一个查询是否比第二个查询运行得更快？我通常会检查执行成本（性能和成本）。

首次质询

SQL Server parse and compile time: 
   CPU time = 0 ms, elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.
SQL Server parse and compile time: 
   CPU time = 0 ms, elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

(931 row(s) affected)
Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Visits'. Scan count 2, logical reads 1235, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Assignments'. Scan count 1, logical reads 4657, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

 SQL Server Execution Times:
   CPU time = 203 ms,  elapsed time = 198 ms.

第二次查询

SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.
SQL Server parse and compile time: 
   CPU time = 0 ms, elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

(1 row(s) affected)
Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Visits'. Scan count 1, logical reads 19335, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Assignments'. Scan count 1, logical reads 4657, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

 SQL Server Execution Times:
   CPU time = 109 ms,  elapsed time = 250 ms.

Answer 1

根据统计数据输出，第一个更快。在那里寻找的东西包括经过的时间和读/写的数量。

第一次查询：

SQL Server Execution Times: CPU time = 203 ms, elapsed time = 198 ms.

第二次查询：

SQL Server Execution Times: CPU time = 109 ms, elapsed time = 250 ms.

第一个查询的执行速度稍快一些。毫秒可能看起来微不足道，但考虑到每天运行数千次的查询，所有这些微小的秒数将开始快速加起来。

要寻找的另一件事是读取和写入（这将代表为获取结​​果集而必须完成的工作量。

第一次查询：

(931 row(s) affected) Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. Table 'Visits'. Scan count 2, **logical reads 1235**, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. Table 'Assignments'. Scan count 1, logical reads 4657, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

第二次查询：

(1 row(s) affected) Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. Table 'Visits'. Scan count 1, **logical reads 19335**, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. Table 'Assignments'. Scan count 1, logical reads 4657, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

在这种情况下，

Logical reads表示从缓冲区缓存中检索的数据，这比从physical reads便宜，其中数据是从物理磁盘中检索的（这是昂贵的）。不过，越少越好。

在这种情况下，您的第一个查询在较短的时间内运行（经过时间= 198ms与250ms），并且需要的读取次数少于第二个查询，（1235 vs {{1因此，优先于第二个。

执行计划中的另一个提示是，更粗的条表示传递的数据越多。然而，虽然他们可以给你一个直观的线索，但一定要验证它的结果，因为我经历过显示的执行计划不准确的情况。