檢查基於SQL Server中列條件

Question

如何解決這個問題，在SQL Server

Table: emp 

Pid | Address   | City | datetime     | Edate  | level 
1 | Homeless   | Chen | 2014-11-13 09:32:14.000 |2013-02-10 |3 
1 | 3913 W. Strong | Chen | 2011-03-044 19:04:10.000 |2014-02-04 |7 
1 | 1100 W MALLON  | Chen | 2014-11-13 09:32:14.000 |2013-02-10 |5 
2 | 610 W GARLAND #3 | Hyd | 2013-11-13 09:32:14.000 |2014-04-02 |4 
3 | banvanu   | chen | 2015-03-044 06:04:10.000 |2015-05-06 |6 
3 | naneku   | chen | 2015-03-044 06:04:10.000 |2015-06-09 |4 

基於上面的表格我想輸出類似下面

Pid | Address   | City | datetime     | Edate  | level 
1 | 1100 W MALLON  | Chen | 2014-11-13 09:32:14.000 |2013-02-10 |5 
2 | 610 W GARLAND #3 | Hyd | 2013-11-13 09:32:14.000 |2014-04-02 |4 
3 | naneku   | chen | 2015-03-044 06:04:10.000 |2015-06-09 |4 

我們需要從獲取地址，城市基於以下條件的同一表

我們得到輸出的條件很少：第一級根據pid檢查max（datetime）如果max（datetime）的值對於相同的pid相同，則相同的pid需要檢查max（edat E）如果我們再相同值，那麼我們需要檢查最大（級）particilar患者檢索地址，城市對於PID

我試着像下面

select * from (select *,row_number()over(partition by id ,order by datetime,edate,level)as rno 
       from emp) 
where rno=1 

，但上面的查詢沒有給出預期導致 請告訴我如何寫查詢才達到在SQL Server這個任務

Answer 1

你需要在窗口函數中使用降序排列：

select [Pid], [Address], [City], [datetime], [Edate], [level] 
from (
    select * 
     , rn= row_number() over (partition by [pid] 
           order by [datetime] desc, 
             [edate] desc, 
             [level] desc 
           ) 
    from emp 
) a 
where rn = 1; 

Answer 2

試試這個：

WITH Empgroup AS (
    SELECT 
     Pid 
     , [Address] 
     , City 
     , [DateTime] 
     , EDate 
     , [Level] 
     , ROW_NUMBER() OVER (PARTITION BY Pid ORDER BY [DateTime] DESC, EDate DESC, [Level] DESC) AS RN 
    FROM 
     Emp 
) 
SELECT 
    Pid 
    , [Address] 
    , City 
    , [DateTime] 
    , EDate 
    , [Level] 
FROM 
    Empgroup 
WHERE 
    RN = 1 

功能區別在於ROW_NUMBER()的排序。然而，我也選擇使用通用表達式來使查詢更清晰。