有条件的每月总和

Question

我有一张桌子，我们称它们为表格 SummaryDATA

NIP  NAME   DEPARTMENT       STATUSIN           STATUSOUT             TOTALLOSTTIME
------------------------------------------------------------------------------------------------
A1   ARIA   BB         2020-01-21 08:06:23  2020-01-21 11:58:36         01:05:00             
A1   ARIA   BB         2020-01-22 07:34:27  2020-01-22 17:19:47         00:30:00               
A1   ARIA   BB         2020-01-23 08:30:00  2020-01-23 11:00:00         00:00:00
A1   ARIA   BB         2020-01-24 08:05:00  2020-01-24 10:30:00         01:00:00
A2   BELLE  BB         2020-01-21 07:06:20  2020-01-21 13:58:31         00:30:00             
A2   BELLE  BB         2020-01-22 07:34:27  2020-01-22 17:19:47         00:15:00               
A2   BELLE  BB         2020-01-23 07:06:00  2020-01-23 10:30:00         00:00:00
A2   BELLE  BB         2020-01-24 09:06:00  2020-01-23 10:30:00         00:05:00
A3   CHLOE  CC         2020-01-21 07:06:23  2020-01-21 11:55:30         00:30:00             
A3   CHLOE  CC         2020-01-22 07:34:27  2020-01-22 17:00:44         01:00:00               
A3   CHLOE  CC         2020-01-23 08:37:00  2020-01-23 11:13:00         00:35:00
A3   CHLOE  CC         2020-01-24 08:09:00  2020-01-24 10:22:00         00:00:00
A4   ZIYA   CC         2020-01-21 07:06:20  2020-01-21 13:58:31         00:30:00             
A4   ZIYA   CC         2020-01-22 07:34:27  2020-01-22 17:19:47         00:15:00               
A4   ZIYA   CC         2020-01-23 06:06:00  2020-01-23 11:30:00         00:45:00
A4   ZIYA   CC         2020-01-24 09:06:00  2020-01-23 15:30:00         00:00:00
A5   BRIAN  BB         2020-01-21 08:06:23  2020-01-21 11:58:36         01:10:00             
A5   BRIAN  BB         2020-01-22 07:34:27  2020-01-22 17:19:47         01:00:00               
A5   BRIAN  BB         2020-01-23 08:30:00  2020-01-23 11:00:00         00:30:00
A5   BRIAN  BB         2020-01-24 08:05:00  2020-01-24 10:30:00         00:10:00

我需要SELECT NIP,NAME,DEPARTMENT,STATUSIN,TOTALLOSTTIME每月和每年where sum(TOTALLOSTTIME) > '02:00:00' 。 那么TOTALLOSTTIME中的数据类型列是nvarchar

这就是我的意思的 output：

    NIP  NAME   DEPARTMENT       STATUSIN           STATUSOUT             TOTALLOSTTIME
    ------------------------------------------------------------------------------------------------
    A1   ARIA   BB         2020-01-21 08:06:23  2020-01-21 11:58:36         01:05:00             
    A1   ARIA   BB         2020-01-22 07:34:27  2020-01-22 17:19:47         00:30:00
    A1   ARIA   BB         2020-01-24 08:05:00  2020-01-24 10:30:00         01:00:00
    A3   CHLOE  CC         2020-01-21 07:06:23  2020-01-21 11:55:30         00:30:00             
    A3   CHLOE  CC         2020-01-22 07:34:27  2020-01-22 17:00:44         01:00:00               
    A3   CHLOE  CC         2020-01-23 08:37:00  2020-01-23 11:13:00         00:35:00 
    A5   BRIAN  BB         2020-01-21 08:06:23  2020-01-21 11:58:36         01:10:00             
    A5   BRIAN  BB         2020-01-22 07:34:27  2020-01-22 17:19:47         01:00:00               
    A5   BRIAN  BB         2020-01-23 08:30:00  2020-01-23 11:00:00         00:30:00
    A5   BRIAN  BB         2020-01-24 08:05:00  2020-01-24 10:30:00         00:10:00

到目前为止，我一直在尝试这个查询：

select NIP,NAME,DEPARTMENT,x.sum_lost_time,x.months,x.years
 from (select MONTH(STATUSIN)  as [months]
               ,YEAR(STATUSIN) as [years]
               ,NIP
               ,NAME
               ,DEPARTMENT         
               ,convert(varchar,dateadd(second,sum(datediff(second,'00:00:00',cast(TOTALLT as time))),0),108) as sum_lost_time
          from SUMMARYDATA b 
      group by MONTH(STATUSIN)
               ,YEAR(STATUSIN)
               ,NIP
               ,NAME
               ,DEPARTMENT
        )x
WHERE x.sum_lost_time > '10:00:00'

Answer 1

您需要属于总损失时间大于 2 小时的组的行。

您可以为此使用 window 函数：

select nip, name, department, statusin, statusin, totallosttime
from (
    select 
        t.*,
        sum(
            left(totallosttime, 2) * 60 * 60 
            + substring(totallosttime, 4, 2) * 60 
            + right(totallosttime, 2)
        ) over(partition by 
            nip, 
            name, 
            department, 
            year(statusin), 
            month(statusin)
        ) totallostseconds
    from mytable t
) t
where totallostseconds >= 2 * 60 * 60

在子查询中，window sum()计算具有相同nip 、 name 、 department 、 year 和 month 的行的总“丢失”持续时间。 我们使用字符串函数来拆分totallosttime并将其转换为秒数。 剩下要做的就是在外部查询中进行过滤。

DB Fiddle 上的演示：

nip | name  | department | statusin   | statusin   | totallosttime
:-- | :---- | :--------- | :--------- | :--------- | :------------
A1  | ARIA  | BB         | 2020-01-21 | 2020-01-21 | 01:05:00     
A1  | ARIA  | BB         | 2020-01-22 | 2020-01-22 | 00:30:00     
A1  | ARIA  | BB         | 2020-01-23 | 2020-01-23 | 00:00:00     
A1  | ARIA  | BB         | 2020-01-24 | 2020-01-24 | 01:00:00     
A3  | CHLOE | CC         | 2020-01-21 | 2020-01-21 | 00:30:00     
A3  | CHLOE | CC         | 2020-01-22 | 2020-01-22 | 01:00:00     
A3  | CHLOE | CC         | 2020-01-23 | 2020-01-23 | 00:35:00     
A3  | CHLOE | CC         | 2020-01-24 | 2020-01-24 | 00:00:00     
A5  | BRIAN | BB         | 2020-01-21 | 2020-01-21 | 01:10:00     
A5  | BRIAN | BB         | 2020-01-22 | 2020-01-22 | 01:00:00     
A5  | BRIAN | BB         | 2020-01-23 | 2020-01-23 | 00:30:00     
A5  | BRIAN | BB         | 2020-01-24 | 2020-01-24 | 00:10:00