使用此示例https://rickosborne.org/blog/2006/01/calculating-overtime-with-pure-sql/，我有以下查询：
SELECT
Employee,
In_Day,
Job_Code2,
Job_Code,
CASE
WHEN (HoursBefore < 40.0)
AND (HoursAfter > 40.0)
THEN
/* straddler */
CASE
WHEN p.i = 0
THEN Hours + (40.0 - HoursAfter)
ELSE HoursAfter - 40.0
END
/* hours adjustment */
ELSE Hours
/* no adjustment */
END AS Hours,
CASE
WHEN (HoursBefore < 40.0)
AND (HoursAfter > 40.0)
THEN
CASE
WHEN p.i = 0
THEN 1.0
ELSE 1.5
END
/* straddlers */
WHEN HoursBefore >= 40.0
THEN 1.5
/* non-straddlers after 40 hours */
ELSE 1
/* non-straddlers before 40 hours */
END AS Rate,
HoursBefore,
HoursAfter
FROM
(
SELECT
a.Employee,
a.In_Day,
a.Job_Code2,
a.Job_Code,
COALESCE(SUM(b.Hours),0) AS HoursBefore,
a.Hours,
COALESCE(SUM(b.Hours),0) + a.Hours AS HoursAfter
FROM
v_TimeSheet a
LEFT OUTER JOIN
v_TimeSheet b
ON
(
a.Employee = b.Employee)
AND (
b.In_Day < a.In_Day)
GROUP BY
a.Employee,
a.In_Day,
a.Job_Code,
a.Hours ) AS c
INNER JOIN
Pivot p
ON
(
i <
CASE
WHEN (HoursBefore < 40.0)
AND (HoursAfter > 40.0)
THEN 2
ELSE 1
END)
ORDER BY
Employee,
In_Day,
Job_Code;
我在计算每天的小时数时遇到过一个问题。结果适用于桌面上90％的员工，但它似乎并没有逐个加起来。
这是它当前所做的一个示例输出（为我的原因稍作修改）：
Employee In_Day Job_Code2 Job_Code Hours Rate HoursBefore HoursAfter
EMPLOYEE 1 3/16/2015 13041 90100 5 1 0 5
EMPLOYEE 1 3/16/2015 13041 90400 5 1 0 5
EMPLOYEE 1 3/17/2015 13041 90100 4 1 10 14
EMPLOYEE 1 3/17/2015 13041 90400 2 1 10 12
EMPLOYEE 1 3/17/2015 13041 90500 2.5 1 10 12.5
EMPLOYEE 1 3/17/2015 13041 300100 2.5 1 10 12.5
EMPLOYEE 1 3/18/2015 13041 90100 4 1 21 25
EMPLOYEE 1 3/18/2015 13041 90400 3 1 21 24
EMPLOYEE 1 3/18/2015 13041 90500 4 1 21 25
EMPLOYEE 1 3/19/2015 13041 90100 2 1 32 34
EMPLOYEE 1 3/19/2015 13041 90400 2 1 32 34
EMPLOYEE 1 3/19/2015 13041 90500 4 1 32 36
EMPLOYEE 1 3/19/2015 13041 300100 3 1 32 35
EMPLOYEE 1 3/20/2015 13041 90100 2 1.5 43 45
EMPLOYEE 1 3/20/2015 13041 90400 4 1.5 43 47
EMPLOYEE 1 3/20/2015 13041 90500 4 1.5 43 47
EMPLOYEE 1 3/21/2015 13041 90100 3 1.5 53 56
EMPLOYEE 1 3/21/2015 13041 90400 3 1.5 53 56
EMPLOYEE 1 3/21/2015 13041 90500 2 1.5 53 55
EMPLOYEE 1 3/21/2015 13041 300100 2 1.5 53 55
EMPLOYEE 2 3/16/2015 14071 100A19 4.5 1 0 4.5
EMPLOYEE 2 3/16/2015 14071 100A23 5 1 0 5
EMPLOYEE 2 3/17/2015 14071 100A19 2.5 1 9.5 12
EMPLOYEE 2 3/17/2015 14071 100A23 7 1 9.5 16.5
EMPLOYEE 2 3/18/2015 14071 100A19 3 1 19 22
EMPLOYEE 2 3/18/2015 14071 100A24 5.5 1 19 24.5
EMPLOYEE 2 3/20/2015 14071 100A24 9.5 1 27.5 37
EMPLOYEE 2 3/21/2015 14071 100A24 3 1 37 46.5
EMPLOYEE 2 3/21/2015 14071 100A24 6.5 1.5 37 46.5
对于EMPLOYEE 2，通过将3/21记录分成2并且仅有6.5小时的加班时间，可以正确计算所有内容。然而，在员工1上，由于3/19 Job_Code 300100被计算为常规时间而不是加班，他的加班时间为3小时。
我尝试了许多不同的分组方法，看看能否得到我需要的结果。
有没有人建议纠正这个问题，或者有其他解决办法？