Split week based on weightage - sql

Question

I have a weights defined like below in a table.
DayNum | Day | Weight | Cumulative Weight
1 | MON | 0.3 | 0.3
2 | TUE | 0.15 | 0.45 (Sum of Mon and Tues)
3 | WED | 0.1 | 0.55 (Sum of Mon and Tues and Wed)
4 | THU | 0.1 | 0.65
5 | FRI | 0.15 | 0.8
6 | SAT | 0.2 | 1
And I have amounts in another table defined at weekly level (Mon - Sun) like below.
Item | Date | Amount
A | 30-May-16 | 10 ---- Week in May and June
A | 6-Jun-16 | 20
A | 13-Jun-16 | 30 and so on
A | 27-Jun-16 | 60 ---- Week in Jun and July
Now I want to insert into another table at a daily level, for the weeks which are overlapping between 2 different months (in above example - 30 May to 5 Jun).
Can anyone explain how I can achieve this in Oracle.
Output should be like below.
Item | Date | Amount
A | 30-May-16 | 4.5 (2 days from May which are Mon and Tues - so calculation is 10 * 0.45)
A | 1-Jun-16 | 5.5 (5 days from May which is the rest of the week - 10 minus 4.5)
A | 6-Jun-16 | 20 and so on
A | 27-Jun-16 | 39 (4 days from June which are Mon till Thurs - so calculation is 60 * 0.65)
A | 1-Jul-16 | 21 (3 days from July which is the rest of the week - 60 minus 39)

Answer 1

Try:
WITH some_data AS(
select a.*,
trunc( trunc( add_months( "DATE", 1 ), 'MM' ) - "DATE" )
As days_to_end_of_month,
trunc( add_months( "DATE", 1 ), 'MM' )
As start_of_next_month
from amounts a
), some_other_data AS (
SELECT some_data.*,
CASE WHEN days_to_end_of_month >= 6 THEN Amount
ELSE ( SELECT some_data.amount * "Cumulative Weight" FROM Weights w
WHERE some_data.days_to_end_of_month = DayNum )
END as new_Amount
FROM some_data
)
SELECT Item, "DATE", New_Amount as amount
FROM some_other_data
UNION ALL
SELECT Item, start_of_next_month, amount-new_amount
FROM some_other_data
WHERE days_to_end_of_month < 6
ORDER BY "DATE"
;

Answer 2

Oracle Setup:
CREATE TABLE Weights ( DayNum, Day, Weight ) AS
SELECT 1, 'MON', 0.3 FROM DUAL UNION ALL
SELECT 2, 'TUE', 0.15 FROM DUAL UNION ALL
SELECT 3, 'WED', 0.1 FROM DUAL UNION ALL
SELECT 4, 'THU', 0.1 FROM DUAL UNION ALL
SELECT 5, 'FRI', 0.15 FROM DUAL UNION ALL
SELECT 6, 'SAT', 0.2 FROM DUAL;
CREATE TABLE weekly_levels ( Item, "Date", Amount ) AS
SELECT 'A', DATE '2016-05-30', 10 FROM DUAL UNION ALL
SELECT 'A', DATE '2016-06-06', 20 FROM DUAL UNION ALL
SELECT 'A', DATE '2016-06-13', 30 FROM DUAL UNION ALL
SELECT 'A', DATE '2016-06-27', 60 FROM DUAL;
Query:
SELECT item,
start_date,
SUM( amount * weight ) AS amount
FROM (
SELECT item,
"Date" AS start_date,
LEAST( "Date" + INTERVAL '6' DAY, LAST_DAY( "Date" ) ) AS end_date,
amount
FROM weekly_levels
UNION
SELECT item,
GREATEST( "Date", TRUNC( "Date" + INTERVAL '6' DAY, 'MM' ) ) AS start_date,
"Date" + INTERVAL '6' DAY AS end_date,
amount
FROM weekly_levels
) d
INNER JOIN
Weights w
ON ( w.DayNum BETWEEN TO_CHAR( start_date, 'D' )
AND TO_CHAR( end_date, 'D' ) )
GROUP BY item, start_date
ORDER BY item, start_date;
Output:
ITEM START_DATE AMOUNT
---- ------------------- ----------
A 2016-05-30 00:00:00 4.5
A 2016-06-01 00:00:00 5.5
A 2016-06-06 00:00:00 20
A 2016-06-13 00:00:00 30
A 2016-06-27 00:00:00 39
A 2016-07-01 00:00:00 21