SQL Server组按每小时的DateTime计数？

Question

    create table #Events
(
    EventID int identity primary key,
    StartDate datetime not null,
    EndDate datetime not null
)
go
insert into #Events (StartDate, EndDate)
select '2007-01-01 12:44:12 AM', '2007-01-01 12:45:34 AM' union all
select '2007-01-01 12:45:12 AM', '2007-01-01 12:46:34 AM' union all
select '2007-01-01 12:46:12 AM', '2007-01-01 12:47:34 AM' union all
select '2007-01-02 5:01:08 AM', '2007-01-02 5:05:37 AM' union all
select '2007-01-02 5:50:08 AM', '2007-01-02 5:55:59 AM' union all
select '2007-01-03 4:34:12 AM', '2007-01-03 4:55:18 AM' union all
select '2007-01-07 3:12:23 AM', '2007-01-07 3:52:25 AM'

(向http://www.sqlteam.com/article/working-with-time-spans-and-durations-in-sql-server道歉,以获取他们的基础sql)

我试图找到一小时内发生的事件计数,因此结果集看起来像这样:

2007-01-01      12:00     3
2007-01-02       5:00     2
2007-01-03       4:00     1
2007-01-07       3:00     1

我一直在玩dateadd和round并分组,但没有得到它.有人可以帮忙吗？

谢谢.

Answer 1

这个怎么样？假设SQL Server 2008:

SELECT CAST(StartDate as date) AS ForDate,
       DATEPART(hour,StartDate) AS OnHour,
       COUNT(*) AS Totals
FROM #Events
GROUP BY CAST(StartDate as date),
       DATEPART(hour,StartDate)

对于2008年之前:

SELECT DATEADD(day,datediff(day,0,StartDate),0)   AS ForDate,
       DATEPART(hour,StartDate) AS OnHour,
       COUNT(*) AS Totals
FROM #Events
GROUP BY CAST(StartDate as date),
       DATEPART(hour,StartDate)

这导致:

ForDate                 | OnHour | Totals
-----------------------------------------
2011-08-09 00:00:00.000     12       3

Answer 2

或者,只是GROUP BY时间和日期:

SELECT  CAST(Startdate as DATE) as 'StartDate', 
        CAST(DATEPART(Hour, StartDate) as varchar) + ':00' as 'Hour', 
        COUNT(*) as 'Ct'
FROM #Events
GROUP BY CAST(Startdate as DATE), DATEPART(Hour, StartDate)
ORDER BY CAST(Startdate as DATE) ASC

输出:

StartDate   Hour    Ct
2007-01-01  0:00    3
2007-01-02  5:00    2
2007-01-03  4:00    1
2007-01-07  3:00    1

Answer 3

我在其他地方找到了这个.我喜欢这个答案!

SELECT [Hourly], COUNT(*) as [Count]
  FROM 
 (SELECT dateadd(hh, datediff(hh, '20010101', [date_created]), '20010101') as [Hourly]
    FROM table) idat
 GROUP BY [Hourly]