Question

正在运行SQL Server 2012。考虑这个值集：

ID    Status
------------
11    0
11    1
11    2
12    0
12    1
12    2
13    1
14    2

如何检索在ID上分组的交叉状态值？

即。由于ID=13只有Status=1且ID=14只有Status=2，因此上述数据会导致空行：

Status
------

但是，如果我们删除ID=14，结果将是：

Status
------
1

如果我们删除ID=13，结果将是：

Status
------
0
1
2

更新

对于具有任意数量的ID值的任意数量的Status，需要解决该问题。因此，两个SELECT查询是不够的（即使它可能适用于上面的测试数据）。

Answer 1

这是relational division问题。

如果除以空集，则使用双NOT EXISTS方法返回all。

DECLARE @S TABLE (
  [ID] INT PRIMARY KEY)

INSERT INTO @S
VALUES      (11),
            (12)

SELECT DISTINCT [Status]
FROM   YourTable Y1
WHERE  NOT EXISTS(SELECT *
                  FROM   @S
                  WHERE  NOT EXISTS (SELECT *
                                     FROM   YourTable Y2
                                     WHERE  Y1.[Status] = Y2.[Status]
                                            AND Y2.ID = [@S].ID))

Answer 2

此选择适用于任何值：

select distinct status
from a a1
where (select count(distinct id) from a) = (select count(distinct id) from a a2 where a2.status = a1.status)

这是一个完整的脚本，可以处理您问题中的示例数字： create table a（id int，status int）

insert a values (11, 0)
insert a values (11, 1)
insert a values (11, 2)
insert a values (12, 0)
insert a values (12, 1)
insert a values (12, 2)
insert a values (13, 1)
insert a values (14, 2)

select distinct status
from a a1
where (select count(distinct id) from a) = (select count(distinct id) from a a2 where a2.status = a1.status)

delete a where id = 14

select distinct status
from a a1
where (select count(distinct id) from a) = (select count(distinct id) from a a2 where a2.status = a1.status)

delete a where id = 13

select distinct status
from a a1
where (select count(distinct id) from a) = (select count(distinct id) from a a2 where a2.status = a1.status)

Answer 3

我认为最简单（也是最可读）的就是这样做：

select t.status
from Table1 as t
group by t.status
having count(distinct t.id) = (select count(distinct t2.id) from Table1 as t2)

<强> sql fiddle demo

从重复组中查找相交值？

更新

3 个答案: