SQL query to use group by to get the sum of two different columns within a date range - sql

Question

I have two tables time track and absence for an employee.
person_number Measure start_Date end_date Time_type
73636 10 01-Jan-2020 02-Jan-2020 Double
73636 24 06-Jan-2020 08-jan-2020 Double
73636 10 15-Jan-2020 25-Jan-2020 Regular Pay
73636 11.9 06-Jan-2020 08-jan-2020 Double
73636 27 10-Jan-2020 15-Jan-2020 Regular Pay
Absence det
person_number start_Date end_date duration Absence_type
73636 05-Jan-2020 10-Jan-2020 10 Vacation
73636 06-Jan-2020 18-jan-2020 9 Paid Leave
73636 20-Jan-2020 21-jan-2020 1 Paid Leave
Now when i pass the from and to date as 01-Jan-2020 and 31-Jan-2020, the output should look like -
Person_Number Double Regular Hour_code hour_amount
73636 31.9 37 Paid Leave 10
The hour_code should have only "Paid Leave" and no other absences
Now I have written the below query for this
SELECT
distinct person_number,
sum(
CASE
WHEN elements = 'Double' THEN measure
END
) AS OT_Hours,
sum(
CASE
WHEN elements LIKE 'Regular Pay%' THEN measure
END
) AS regular_measure_hours,
sum(
CASE
WHEN absence_name IN ('Paid Leave') THEN absence_duration
END
) AS hour3_amount,
max(
CASE
WHEN absence_name IN ('Paid Leave') THEN 'Paid Leave'
END
) AS hour3_code
FROM
(
select
person_number,
Time_type elements,
Absence_type absence_name,
duration,
measure
from
time_track_tab,
abs_tab,
per_all_people_F papf
where
time_track_tab.person_id = abs_tab.person_id
and abs_tab.person_id = papf.person_id
and abs_tab.Absence_type = 'Paid Leave'
)
group by
person_number
This is giving me multiple row output and calculation of sum is not coming correctly as in between the to and from date there are different dates present for both absence and time track.
My requirement is to calculate the sum of ALL the duration and measure column within these parameter dates. How can i tweak my query to get the correct sum between these dates ?
Is there a way to use partition by or group by or anything else to calculation these correctly in the column

Answer 1

You probably need to group both tables first then join them together to avoid the cross join.
select person_number, TimeTrack.DoublePay, TimeTrack.Regular,
Absenses.Hour_code, Absenses.hour_amount from
per_all_people_F papf,
(select
person_id, sum(duration) as hour_amount, Absence_type as Hour_code
from
abs_tab
where
abs_tab.Absence_type = 'Paid Leave'
and
start_Date between '2020-01-01' and '2020-01-31'
group by person_id,Absence_type
) Absenses,
(select
person_id,
sum(case when Time_type = 'Double' then Measure end) as DoublePay,
sum(case when Time_type = 'Regular Pay' then Measure end) as Regular
from time_track_tab
where
start_Date between '2020-01-01' and '2020-01-31'
group by person_id
) TimeTrack
where
papf.person_id = TimeTrack.person_id
and
papf.person_id = Absenses.person_id
and
papf.person_id = 73636
I made a SqlFiddle if you want to play with it
http://sqlfiddle.com/#!9/03e460/36
Also my 2 cents; I'd recommend left outer joining from the per_all_people_F table or else people without absenses will get filtered out.

Answer 2

See if, what you need is something like this:
select * from
(SELECT person_number,
sum(
CASE
WHEN Time_type = 'Double' THEN measure
END
) AS Double,
sum(
CASE
WHEN Time_type = ('Regular Pay') THEN measure
END
) AS regular
from time_track_tab
group by person_number
) A
inner join
(SELECT
person_number,
sum(
CASE
WHEN Absence_type = 'Vacation' THEN duration
END
) AS Vacation,
sum(
CASE
WHEN Absence_type = ('Paid Leave') THEN duration
END
) AS paidLeave
from abs_tab
group by person_number
)B on A.person_number = B.person_number
here the fiddle:
http://sqlfiddle.com/#!4/21253/2