Question

我有一个如下表所示的记录，例如，同一客户有很多记录，但他们的EffectiveDate和RunDate可能相同或不同。 EffectiveDate和RunDate也可能包括TimeStamp，如2014-02-09 14：15：30.000，并且时间戳可能在同一日期有所不同。

ID  Client        EffectiveDate                 RunDate         
1   abc      01-01-2014 14:15:30.000    09-02-2014 14:15:30.000        
2   abc      02-01-2014 14:15:30.000    10-01-2014 14:15:30.000       
3   abc      02-01-2014 15:19:30.000    09-01-2014 15:19:30.000          
4   abc      03-01-2014 14:15:30.000    10-01-2014 14:15:30.000            
5   abc      03-01-2014 15:20:30.000    09-01-2014 15:20:30.000           
6   abc      02-01-2014 14:15:30.000    10-01-2014 14:15:30.000          
7   abc      04-01-2014 14:30:30.000    10-01-2014 14:30:30.000           
8   abc      04-01-2014 17:15:30.000    10-01-2014 17:15:30.000

我需要一个sql查询，其结果与相同的EffectiveDate相同，它将计算该客户端的记录数，而不管EffectiveDate的Timestamp和RunDate（包括其Timestamp）。我需要像

这样的结果

 ID  Client      EffectiveDate   Number Of Records    
 1   abc          01-01-2014            1
 2   abc          02-01-2014            3
 4   abc          03-01-2014            2
 7   abc          04-01-2014            2

我知道可以在sql查询中使用Group By但是如果在日期中有时间戳，那么我无法获得所需的结果。

Answer 1

使用以下查询获取结果：

    SELECT ID, Client, EffectiveDate, RunDate
    FROM 
        (SELECT ROW_NUMBER() OVER (PARTITION BY A.Client, B.EffectiveDate, 
                B.NumberOfRecords ORDER BY A.ID) AS RowNum, 
                 A.ID, A.Client, B.EffectiveDate, CONVERT(VARCHAR(10), A.RunDate ,105) 
                AS RunDate, B.NumberOfRecords
         FROM #Temp A 
         INNER JOIN (SELECT Client, CONVERT(VARCHAR(10),EffectiveDate ,105) AS 
                          EffectiveDate, COUNT(*) AS NumberOfRecords
                     FROM #Temp
                      GROUP BY Client, CONVERT(VARCHAR(10),EffectiveDate ,105)) B 
         ON A.Client = B.Client AND CONVERT(VARCHAR(10), A.EffectiveDate ,105) = 
            B.EffectiveDate) AS C
    WHERE RowNum = 1

Answer 2

您需要convert datetime to date删除timestamp

中的group by

select  Client,convert(date,EffectiveDate), count(1) [Number Of Records]
from tablename 
group by Client,convert(date,EffectiveDate)

如果你想要Id的结果，那么试试这个（未经测试）。

;WITH cte
     AS (SELECT Row_number() OVER(partition BY CONVERT(DATE, EffectiveDate) ORDER BY id) rn,*
         FROM   tablename)
SELECT a.ID,
       a.Client,
       a.EffectiveDate,
       b.[Number Of Records]
FROM   cte a
       JOIN(SELECT Client,
                   CONVERT(DATE, EffectiveDate) EffectiveDate,
                   Count(1)                     [Number Of Records]
            FROM   tablename
            GROUP  BY Client,
                      CONVERT(DATE, EffectiveDate)) b
         ON CONVERT(DATE, a.EffectiveDate) = b.EffectiveDate 
where rn =1

Answer 3

像这样选择日期部分

select convert(varchar,getdate(),105)

也用相同的

分组

Answer 4

select min(id) id,client,trunc(effectivedate) effectivedate,
min(trunc(rundate)) rundate,
count(client) no_of_records 
from tempd
group by client,trunc(effectivedate)
order by 1

你可以尝试这个在oracle sql中运行良好和快速的工作，对于rundate，取决于你想要保持最小或最大的一个;在这个查询中我保持最小如果你想要最大然后把max而不是min，你会得到你想要的结果。

Answer 5

尝试以下查询：

SELECT Min(ID) AS ID, 
    Client, 
    CONVERT(DATE,EffectiveDate) AS EffectiveDate, 
    MIN(RunDate) AS RunDate, 
    COUNT(*) AS [Number Of Records]
FROM ClientTableName
GROUP BY Client, CONVERT(Date,EffectiveDate)

注意：唯一的区别是返回的日期格式。这可以很容易地改为所需的格式

使用Group By子句的SQL查询结果，不包括日期字段中的时间戳

5 个答案: