Sql server SQL-大表和小表之间的内部联接，带有筛选器

我有一个简单的SQL查询，它在表a和表B之间进行内部连接，并进行过滤。表A非常大，而表B相当小。我只想返回与某个筛选器匹配的行，例如，table_B.some_col IN（'a'，'B'，'c'，'d'）。当然，我把这个过滤器作为WHERE子句。但是，建议我复制相同的筛选器，但要复制表A并将其放入联接中。比如说,

SELECT
*
FROM table_A
INNER JOIN table_B
    ON table_A.pk = table_B.pk
    AND table_A.some_col IN ('a', 'b', 'c', 'd')
WHERE
table_B.some_col IN ('a', 'b', 'c', 'd')

---

据称，这样做可以改进查询计划。不幸的是，由于机器上的安全限制，我无法检查查询计划。这种说法是真的吗？

在

和
或
的Where
子句中添加一个过滤器没有任何区别

这里有一些测试

使用的查询..

从中选择*
命令
参加
人力资源部员工
关于h.empid=o.empid

在And子句中筛选..

empid=4

执行计划：


设置统计IO:

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

在Where子句处筛选：

其中empid=4

执行计划：


统计IO:

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

在每种情况下，SQLServer都足够聪明，可以重新安排查询（因为它是内部联接），只获取相关行
在
和
或
处添加一个过滤器，其中
子句没有任何区别

这里有一些测试

使用的查询..

从中选择*
命令
参加
人力资源部员工
关于h.empid=o.empid

在And子句中筛选..

empid=4

执行计划：


设置统计IO:

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

在Where子句处筛选：

其中empid=4

执行计划：


统计IO:

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

Table 'orders'. Scan count 1, logical reads 20, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Employees'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

在每种情况下，SQLServer都足够聪明，可以重新排列查询（因为它是内部联接），只获取相关行
等等，您的两个查询实际上是不同的查询，没有表A。某些列在。。。您可以从B中选择任何在A中具有匹配主键的选项-除非B中的某些列在A中是重复值，否则您的两个查询实际上是不同的查询，没有表A。。。除非B中的某个列是A中的重复值，否则您可以选择从B中选择的、在A中具有匹配主键的任何列