Question

我对SQL很新。这是我的最终目标：

2015年12月31日之后由护理人员按日历日工作的小时数

这是我现在的代码

select br_ID, car_ID, car_FirstName, car_LastName, car_SkillCd, car_Role, car_EmpSubCod_UD, schst_Name, svcc_Name, dbo.ConvertTimeFromUtc(sch_StartTime , tz_Data), dbo.ConvertTimeFromUtc(sch_endtime , tz_Data), sch_Units, sch_PayDate, sch_PayUnits, cli_LastName, cli_FirstName 
from T_Schedules
inner join T_Branches on br_ID = sch_BranchID
inner join T_TimeZones on tz_ID = br_TimeZoneID
Full outer join T_Caregiver ON T_schedules.sch_CaregiverID = T_Caregiver.car_ID
full outer join T_ServiceCode on T_Schedules.sch_ServiceCodeID = T_ServiceCode.svcc_ID
full outer join T_ScheduleStatus on T_schedules.sch_ScheduleStatusID = T_ScheduleStatus.schst_ID
full outer join T_Client on T_Schedules.sch_ClientID = T_Client.cli_ID
where sch_ScheduleStatusID = 3 and dbo.ConvertTimeFromUtc(sch_StartTime , tz_Data) >= '2016-01-01'

任何人都知道如何按照照顾者分组。我不能只计算预定的单位，因为如果它是一夜之间的班次，那么这些时间可以在2个日历日之间分配。

Answer 1

隔夜转移是一个问题，很可能会扭曲你的结果。除非你回去并将1/2转换为一天和1/2到另一天，否则SQL无法为你做任何事情。（我认为）我将如何处理这个问题是一个简单的小组，看看它是否足够好：

SELECT dbo.ConvertTimeFromUtc(sch_StartTime , tz_Data), 
br_ID, car_ID, car_FirstName, 
car_LastName, car_SkillCd, car_Role, 
car_EmpSubCod_UD, schst_Name, svcc_Name, 
dbo.ConvertTimeFromUtc(sch_endtime , tz_Data), 
SUM(sch_Units), 
sch_PayDate, sch_PayUnits, cli_LastName, cli_FirstName 
FROM T_Schedules
INNER JOIN T_Branches on br_ID = sch_BranchID
INNER JOIN T_TimeZones on tz_ID = br_TimeZoneID
FULL OUTER JOIN T_Caregiver ON T_schedules.sch_CaregiverID = T_Caregiver.car_ID
FULL OUTER JOIN T_ServiceCode on T_Schedules.sch_ServiceCodeID = T_ServiceCode.svcc_ID
FULL OUTER JOIN T_ScheduleStatus on T_schedules.sch_ScheduleStatusID = T_ScheduleStatus.schst_ID
FULL OUTER JOIN T_Client on T_Schedules.sch_ClientID = T_Client.cli_ID

WHERE sch_ScheduleStatusID = 3 and dbo.ConvertTimeFromUtc(sch_StartTime , tz_Data) >= '2016-01-01'

GROUP BY [sch_StartTime]

Answer 2

问题：从sch_StartTime到sch_endtime的转变可能会在午夜过后，所以一天工作一小时，下一天工作一些。

我们写了两个查询：一个用于第一天，一个用于下一个，然后将两个集合与UNION ALL粘合在一起。以下是使用WITH子句逐步构建的查询：

with worked as
(
  select 
    sch_caregiverid as caregiverid,
    dbo.ConvertTimeFromUtc(sch_StartTime, tz_Data) as starttime,
    dbo.ConvertTimeFromUtc(sch_endtime, tz_Data) as endtime
  from T_Schedules
  where sch_ScheduleStatusID = 3 and dbo.ConvertTimeFromUtc(sch_StartTime , tz_Data) >= '2016-01-01'
)
, day1 as
(
  select
    caregiverid,
    starttime, 
    case when day(starttime) <> day(endtime) then 
      datetimefromparts(year(starttime), month(starttime), day(starttime), 23, 59, 59, 999)
    else 
      endtime
    end as endtime
  from worked
)
, day2 as
(
  select 
    caregiverid,
    datetimefromparts(year(endtime), month(endtime), day(endtime), 0, 0, 0, 0) as starttime, 
    endtime
  from worked
  where day(starttime) <> day(endtime)
)
, bothdays as
(
  select caregiverid, starttime, endtime from day1
  union all
  select caregiverid, starttime, endtime from day2
)
, hours_worked as
(
  select 
    caregiverid, 
    convert(date, starttime) as whichday, 
    sum(datediff(hour, starttime, endtime)) as total
  from bothdays
  group by caregiverid, convert(date, starttime)
)
select *
from t_caregiver 
join hours_worked on hours_worked.caregiverid = t_caregiver.car_id;

用于组的SQL编码和具有多个表和where语句的sum

2 个答案: