Question

假设两个表具有以下内容：

insert into tmp1 values(12345) // orderid
insert into tmp1 values(54321)
insert into tmp1 values(77777)

insert into tmp2 values(12345, 1) // orderid, code
insert into tmp2 values(12345, 2)
insert into tmp2 values(54321, 1)
insert into tmp2 values(77777, 1)
insert into tmp2 values(77777, 2)
insert into tmp2 values(77777, 3)

然后我写下以下内容：

select a.orderid from tmp1 a
inner join tmp2 b on b.orderid = a.orderid where b.code in (1,2)

有没有办法可以获得输出：

12345,
54321,
77777

这是一个报告，我只想知道由于内部联接而没有复制的订单号，尽管内部联接很重要，因为可以选择多个代码。

Answer 1

SELECT DISTINCT a.orderid 
FROM   tmp1 a 
       INNER JOIN tmp2 b 
               ON b.orderid = a.orderid 
WHERE  b.code IN ( 1, 2 )

Answer 2

select distinct a.orderid from tmp1 a inner join tmp2 b
on b.orderid = a.orderid where b.code in (1,2)

Answer 3

您可以像这样使用UNION运算符 -

Select orderid from tmp1
Union 
Select orderid from tmp2 Where code In (1,2);

或者你也可以使用不同的内部联接 -

SELECT 
    DISTINCT t1.Orderid
FROM tmp1 t1
INNER JOIN tmp2 t2 
ON t1.orderid = t2.orderid
WHERE t2.code IN (1, 2);

Answer 4

避免任何加入：

select distinct t2.orderid
from temp2 t2
where t2.code in (1,2)

使用存在（“半连接”）：

select t1.orderid
from tmp1 t1
    where exists (
    select null
    from temp2 t2
    where t1.orderid = t2.orderid and t2.code in (1,2)
    )

使用窗口函数ROW_NUMBER（）：

select d.* from (
    select t2.orderid
         , row_number() over(partition by t2.orderid order by t2.code) rn
    from temp2 t2
    where t2.code in (1,2)
    ) d
where d.rn = 1

对于这个简单的例子，这第三个选项似乎有些过分，但该技术对于查找诸如“第一”或“最后”或“最早”或“最新”等信息特别有用。

如何在T-SQL中模仿LIMIT 1

4 个答案: