Question

如果您想尝试，我会在此SQL Fiddle上构建架构。

我有以下数据

AutoID         Apartment      Flag       Description
======         =========      ====       ===========
1              1              NO         Device 1
2              1              NO         Device 2
3              1              NO         Device 3
4              2              NO         Device 4
5              2              NO         Device 5
6              3              NO         Device 6
7              3              NO         Device 7
8              3              YES        Device 8
9              3              NO         Device 9

我正在尝试使用以下规则获取数据

仅选择不同的公寓值
如果标记为是，则选择该项为不同的值

因此，如果我运行SQL语句，我最终将使用

AutoID         Apartment      Flag       Description
======         =========      ====       ===========
1              1              NO         Device 1
4              2              NO         Device 4
8              3              YES        Device 8

我一直试图用OVER（）和PARTITION BY玩弄运气不好。非常感谢任何帮助或指导。

Answer 1

;WITH CTE AS
(
    SELECT  *,
            RN = ROW_NUMBER() OVER(PARTITION BY Apartment
                                   ORDER BY CASE WHEN [Flag] = 'YES' THEN 1 ELSE 2 END,
                                            [AutoID])
    FROM TableA
)
SELECT *
FROM CTE
WHERE RN = 1

Here is修改后的sqlfiddle

Answer 2

第一步 - 为每个组中的订单号分配相同的公寓号

第二步 - 只选择第一步

select AutoID, Apartment, Flag, Description from (
select AutoID, Apartment, Flag, Description, row_number() over (partition by apartment order by Flag desc) rnum 
from table1)
where rnum = 1

SELECT与其他列值的条件不同

2 个答案: