Question

我使用以下SQL（SQL Server 2005）来计算每列的总时间：

DECLARE @Param1 DATETIME
DECLARE @Param2 DATETIME
DECLARE @Param3 DATETIME

SET @Param1 = '2009-01-01'
SET @Param2 = '2009-09-09'

SELECT  SUM(IdleSec) AS TotalIdleSec, 
        SUM(ProductionSec) AS TotalProductionSec, 
        SUM(UplineSec) AS TotalUplineSec, 
        SUM(DownlineSec) AS TotalDownlineSec, 
        SUM(UserSec) AS TotalUserSec
FROM            Job
WHERE   (DateTime >= dbo.FormatDateTime(@Param1, 'yyyy-mm-dd')) 
        AND 
        (DateTime < dbo.FormatDateTime(@Param2, 'yyyy-mm-dd'))

GO

这会为上面的项目返回一个包含 1行的表，效果很好。

我有点不确定如何以每天/每周/每月的SUM值返回表格即。

日期范围内每个日期的所有值的总和。
日期范围内每个WEEK的所有值的总和。
日期范围内每个MONTH的所有值的总和。

我确信有一种简单的方法可以做到这一点，但不确定自己。我看过一些使用DAY（日期）命令的教程，但我尝试过，似乎无法得到我需要的东西。

我期待着您的出色帮助。

Answer 1

您需要一张日期为

的表格

DECLARE @StartDate DATETIME,
        @EndDate DATETIME

SELECT  @StartDate = '01 Sep 2009',
        @EndDate = '09 Sep 2009'

DECLARE @RunDate DATETIME
SELECT  @RunDate = @StartDate

DECLARE @DayTable TABLE(
        DayDate DATETIME
)

WHILE @RunDate <= @EndDate 
BEGIN
    INSERT INTO @DayTable (DayDate) SELECT @RunDate
    SET @RunDate = @RunDate + 1
END

SELECT * FROM @DayTable

然后将查找表连接到此表，并按日期表中的日期分组。

这应该为您提供每日价值

然后你需要为每周值做同样的事情。这将为您提供所有日子，即使该日期没有统计数据

Answer 2

您可以在查询中添加GROUP BY子句：

MONTH：

SELECT  SUM(IdleSec) AS TotalIdleSec, 
        SUM(ProductionSec) AS TotalProductionSec, 
        SUM(UplineSec) AS TotalUplineSec, 
        SUM(DownlineSec) AS TotalDownlineSec, 
        SUM(UserSec) AS TotalUserSec,
        DATEPART(year, DateTime) as Year,
        DATEPART(month, DateTime) as Month
FROM            Job
WHERE   (DateTime >= dbo.FormatDateTime(@Param1, 'yyyy-mm-dd')) 
        AND 
        (DateTime < dbo.FormatDateTime(@Param2, 'yyyy-mm-dd'))
GROUP BY DATEPART(year, DateTime),
        DATEPART(month, DateTime);

WEEK

：

SELECT  SUM(IdleSec) AS TotalIdleSec, 
        SUM(ProductionSec) AS TotalProductionSec, 
        SUM(UplineSec) AS TotalUplineSec, 
        SUM(DownlineSec) AS TotalDownlineSec, 
        SUM(UserSec) AS TotalUserSec,
        DATEPART(year, DateTime) as Year,
        DATEPART(week, DateTime) as Week
FROM            Job
WHERE   (DateTime >= dbo.FormatDateTime(@Param1, 'yyyy-mm-dd')) 
        AND 
        (DateTime < dbo.FormatDateTime(@Param2, 'yyyy-mm-dd'))
GROUP BY DATEPART(year, DateTime),
        DATEPART(week, DateTime);

by DAY：

 SELECT  SUM(IdleSec) AS TotalIdleSec, 
        SUM(ProductionSec) AS TotalProductionSec, 
        SUM(UplineSec) AS TotalUplineSec, 
        SUM(DownlineSec) AS TotalDownlineSec, 
        SUM(UserSec) AS TotalUserSec,
        DATEPART(year, DateTime) as Year,
        DATEPART(dayofyar, DateTime) as Day
FROM            Job
WHERE   (DateTime >= dbo.FormatDateTime(@Param1, 'yyyy-mm-dd')) 
        AND 
        (DateTime < dbo.FormatDateTime(@Param2, 'yyyy-mm-dd'))
GROUP BY DATEPART(year, DateTime),
        DATEPART(dayofyear, DateTime);

对于DAY，还有其他选项，例如使用CAST提取日期部分或您似乎拥有的FormatDateTime函数。我使用DATEPART来保持一致性。

Answer 3

查看DATEDIFF命令，这可以为您提供两个日期之间的天数/周/月/年等数字。

SQL语法：为一系列日期之间的每个日期选择结果

3 个答案: