Sql 计算日历期间内多个开始-停止记录之间的天数

在给定的时间段内，活动可以启动和停止0次或更多次。每次这样做时，都会生成一个带有日期和操作的记录——我将这些操作称为“激活a”和“停用D”。我需要总结一个月内“活动”天数的总数。我想要一个简单的方法来做到这一点与例外和特殊情况整洁。我必须在tsql中执行此操作

以下是我迄今为止采取的程序方法：

在此期间是否有开始或停止事件？ a如果没有，则查找时间段之前的最新事件 如果该事件已开始，则该期间的所有天都处于活动状态 ii如果事件停止，则该期间没有活动日 三、完成 b如果是的话，那就是程序性检查的老鼠窝，如果是的话。 像往常一样，我希望了解其他人是如何看待这个问题的，而不是仅仅开始编写一个例程，按时间顺序对记录进行排序并循环遍历它们

典型数据集：

6/11/2014  A
6/20/2014  D
6/24/2014  A
6/26/2014  D
6/29/2014  A

这里的关注期是6月，第一个动作是激活，因此活动在6月11日之前一直处于非活动状态。同样地，在6月29日，事情再次活跃起来，因此这个月的剩余时间被视为活跃。在D和a日期上使用datediff函数，我得到9+2，一个月的最后一天总共有12个活动日

同样，仅对于一个周期的数据集，可能有0个或多个记录

当然，数据也可以是这样的：

6/11/2014  D
6/20/2014  A
6/24/2014  D
6/26/2014  A
6/29/2014  D

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select e, count(*) from dates GROUP BY e
OPTION (MAXRECURSION 0)

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select 
sum(case when e='A' THEN 1 ELSE 0 END) as A_days,
sum(case when e='D' THEN 1 ELSE 0 END) as D_days
from dates
OPTION (MAXRECURSION 0)

在这里，该活动一直处于“活动”状态，直到11月6日被停用。简单的日期区分D-A日期给了我额外的4+3天，总共17天

---

我担心如果我选择，排序然后做，而我将得到缠绕轴。我想知道其他人可能会采取什么更圆滑的方法。

此查询将为您提供给定时间段内的所有天数，以及当天的相应事件：

MS SQL Server 2008架构设置：

问题1：

使用上述查询，您将能够轻松生成计数查询，如下所示：

6/11/2014  D
6/20/2014  A
6/24/2014  D
6/26/2014  A
6/29/2014  D

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select e, count(*) from dates GROUP BY e
OPTION (MAXRECURSION 0)

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select 
sum(case when e='A' THEN 1 ELSE 0 END) as A_days,
sum(case when e='D' THEN 1 ELSE 0 END) as D_days
from dates
OPTION (MAXRECURSION 0)

或者像这样：

6/11/2014  D
6/20/2014  A
6/24/2014  D
6/26/2014  A
6/29/2014  D

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select e, count(*) from dates GROUP BY e
OPTION (MAXRECURSION 0)

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select 
sum(case when e='A' THEN 1 ELSE 0 END) as A_days,
sum(case when e='D' THEN 1 ELSE 0 END) as D_days
from dates
OPTION (MAXRECURSION 0)

---

此查询将为您提供给定期间的所有天，以及该天的相应事件：

MS SQL Server 2008架构设置：

问题1：

使用上述查询，您将能够轻松生成计数查询，如下所示：

6/11/2014  D
6/20/2014  A
6/24/2014  D
6/26/2014  A
6/29/2014  D

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select e, count(*) from dates GROUP BY e
OPTION (MAXRECURSION 0)

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select 
sum(case when e='A' THEN 1 ELSE 0 END) as A_days,
sum(case when e='D' THEN 1 ELSE 0 END) as D_days
from dates
OPTION (MAXRECURSION 0)

或者像这样：

6/11/2014  D
6/20/2014  A
6/24/2014  D
6/26/2014  A
6/29/2014  D

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select e, count(*) from dates GROUP BY e
OPTION (MAXRECURSION 0)

declare @start date, @end date
select @start = '2014-06-01', @end = '2014-06-30'
;with dates as (
  SELECT @start AS d, (SELECT TOP 1 CASE WHEN event = 'A' THEN 'D' ELSE 'A' END FROM Table1 ORDER BY dt) AS e
  UNION ALL 
  SELECT dateadd(day, 1, d), COALESCE((SELECT event FROM Table1 WHERE dt = dateadd(day, 1, dates.d)), dates.e)
  FROM dates 
  WHERE dates.d < @end)
select 
sum(case when e='A' THEN 1 ELSE 0 END) as A_days,
sum(case when e='D' THEN 1 ELSE 0 END) as D_days
from dates
OPTION (MAXRECURSION 0)

---

将表1定义为CTE是否合法，例如：表1 d，t为…，日期为[您的上述定义]？我还没有看到关于在同一范围内混合使用非递归和递归定义的参考。我不明白你的问题。好了：我想你可以自己试试……是的，我做了，它完全符合我的需要……工作完美。顺便说一句，如果你注意到我犯了一个错误。你需要自己修复它，非常容易修复。如果您在第1天有一个事件，即如果表中存在@start date的事件，则日期CTE中的初始e值将错误地设置为相反的事件。很容易修复，我将留给你自己。同样，在以下情况下，将表1定义为CTE是否合法：表1 d，t为…，日期为[你的上述定义]？我还没有看到关于在同一范围内混合使用非递归和递归定义的参考。我不明白你的问题。好了：我想你可以自己试试……是的，我做了，它完全符合我的需要……工作完美。顺便说一句，如果你注意到我犯了一个错误。你需要自己修复它，非常容易修复。如果您在第1天有一个事件，即如果表中存在@start date的事件，则日期CTE中的初始e值将错误地设置为相反的事件。很容易修复，由我自己决定