Question

我有这个数据

ID Name State Check
1  Raj  UP     ok
1  Raj  UP     Notok
2  bob  MP     ok
2  bob  MP     ok
3  joy  MH     Notok
3  joy  MH     Notok

我想要关注数据

ID Name State Check  flag
1  Raj  UP     ok      y  
1  Raj  UP     Notok   n 
2  bob  MP     ok      y
2  bob  MP     ok      n
3  joy  MH     Notok   y
3  joy  MH     Notok   n

我想确定哪个被选中。

Answer 1

您应该指定规则，但看起来任何ok都会命名为ok：

select id, name, state,
       (case when sum(case when [check] = 'ok' then 1 else 0 end) > 0
             then 'ok'
             else 'Notok'
        end) as [check]
from t
group by id, name, state;

注意：check是SQL中的关键字，因此它是列的错误名称。在您的情况下，您也可以这样做：

select id, name, state, min(lower([check])) as [check]
from t
group by id, name, state;

此版本使用逻辑值的字母顺序。

Answer 2

create table #Temp (ID int, Name Varchar(8000), State char(2), [Check] varchar(8000))
Insert #Temp Values (1,  'Raj',  'UP',     'ok')
Insert #Temp Values (1  ,'Raj',  'UP',     'Notok')
Insert #Temp Values (2  ,'bob',  'MP',     'ok')
Insert #Temp Values (2  ,'bob',  'MP',     'ok')
Insert #Temp Values (3  ,'joy',  'MH',     'Notok')
Insert #Temp Values (3  ,'joy',  'MH',     'Notok')


;With cteRowsAssigned AS
(
Select  Row_Number() Over (Partition By Name Order By [Check] Desc) RowNum,
        *
    From #Temp
)
Select ID, Name, State, [Check] From cteRowsAssigned Where RowNum = 1 Order By ID

Answer 3

如果您的意思是ok优先于Notok

select id, name, state, case when max(
       case when [check] = 'ok' then 1 else 0 end) = 1
       then 'ok'
       else 'Notok'
    end as [check]
from table
group by id, name, state;

Answer 4

最简单的形式是：

SELECT id,name,state FROM t where <whatever filter you want to set> GROUP BY id,name,state

从基于多个筛选器的SQL Server中选择重复行

4 个答案: