Question

在这里运行SQL Server 2008R2，我有一个问题，我的逻辑还没有完全开始，大声笑

好吧，这就是我所拥有的：

+-----------+------------+------------------+-------+----------+----------+
| DAILYDAYS | MISSED_DAY | MISSED_DAY_COUNT | COUNT |  START   |   END    |
+-----------+------------+------------------+-------+----------+----------+
|  20140114 | (null)     | (null)           |     0 | 20140114 | 20140122 |
|  20140115 | (null)     | (null)           |     1 | 20140114 | 20140122 |
|  20140116 | (null)     | (null)           |     2 | 20140114 | 20140122 |
|  20140117 | (null)     | (null)           |     3 | 20140114 | 20140122 |
|  20140118 | (null)     | (null)           |     4 | 20140114 | 20140122 |
|  20140119 | (null)     | (null)           |     5 | 20140114 | 20140122 |
|  20140120 | 20140120   | 1                |     6 | 20140114 | 20140122 |
|  20140121 | (null)     | (null)           |     7 | 20140114 | 20140122 |
|  20140122 | (null)     | (null)           |     8 | 20140114 | 20140122 |
+-----------+------------+------------------+-------+----------+----------+

这是我需要的地方：

+-----------+------------+------------------+-------+----------+----------+
| DAILYDAYS | MISSED_DAY | MISSED_DAY_COUNT | COUNT |  START   |   END    |
+-----------+------------+------------------+-------+----------+----------+
|  20140114 | (null)     | (null)           |     0 | 20140114 | 20140122 |
|  20140115 | (null)     | (null)           |     1 | 20140114 | 20140122 |
|  20140116 | (null)     | (null)           |     2 | 20140114 | 20140122 |
|  20140117 | (null)     | (null)           |     3 | 20140114 | 20140122 |
|  20140118 | (null)     | (null)           |     4 | 20140114 | 20140122 |
|  20140119 | (null)     | (null)           |     5 | 20140114 | 20140122 |
|  20140120 | 20140120   | 1                |     6 | 20140114 | 20140122 |
|  20140121 | (null)     | 2                |     7 | 20140114 | 20140122 |
|  20140122 | (null)     | 3                |     8 | 20140114 | 20140122 |
+-----------+------------+------------------+-------+----------+----------+

我为你们创建了一个SQL小提琴，我尝试过一个左手连接尝试但是我失败了，因此我在这里问：

http://sqlfiddle.com/#!3/043de/7

问题：

我有两个表：1包含一个表，其中包含每个时段的行，另一个表只包含＆＃34;错过的日期＆＃34;

我需要根据开始和结束[日期]计算错过天数并在上表中保持运行计数

在我的示例中，我的第一个表包含2014-01-14到2014-01-22每天的行数，另一个表格在2014-01-20错过了一天

所以，我需要rank()或dense_rank()或任何其他变体才能获得1,2,3计数......

请注意，在这种情况下我不能使用CTE，因为我的全表超过800,000行，因此不会有效IMO

希望有人能协助

谢谢：）

编辑：我忘了提到重复是没有必要的，它只是我想到计数1,2,3与rank（）的唯一方法是让值一直出现以便做一个＆＃34 ;＆＃34;分区在rank（）函数内部

我已经更新了上面的表格，有没有办法计算错过的天数，直到您能想到的结束日期为止？

Answer 1

我取消了左连接并用工会取而代之。

select 
  [dailydays]
  ,null as [missed_day]
  ,null as [missed_day_count] 
  ,[count]
  ,[start]
  ,[end]
from tbl_1
where dailydays < (select min(missed_day) from tbl_2)
union 
select 
   t1.[dailydays]
  ,t2.[missed_day]
  ,(t1.[count] - (select min([count]) from tbl_1 where [dailydays] = (select min([missed_day]) from tbl_2)) + 1)
  ,t1.[count]
  ,t1.[start]
  ,t1.[end]
from tbl_1 t1
left join tbl_2 t2 on t1.[dailydays] = t2.[missed_day]
where t1.dailydays >= (select min(missed_day) from tbl_2)
order by [dailydays]

修改或者是一个复杂的案例陈述：

select t1.[dailydays] ,t2.[missed_day] , case when (t1.[count] - (select min([count]) from tbl_1 where [dailydays] = (select min([missed_day]) from tbl_2)) + 1) < 0 then 0 else (t1.[count] - (select min([count]) from tbl_1 where [dailydays] = (select min([missed_day]) from tbl_2)) + 1) end as [missed_day_count] ,t1.[count] ,t1.[start] ,t1.[end] from tbl_1 t1 left join tbl_2 t2 on t1.[dailydays] = t2.[missed_day] order by [dailydays]

SQL使用rank（）或其他变体错过了一天的计数

1 个答案: