Adjust column value based on difference between 2 dates

Question

I have sales data at the start of every month in a quarter.

Example: For 22-Q1 quarter, I have sales on 3 dates (1st Jan, 1st Feb and 1st March)

Date	Country	Region	Sales
01/01/2022	UK	EMEA	100,000
02/01/2022	UK	EMEA	170,000
03/01/2022	UK	EMEA	120,000
01/01/2022	US	AMS	90,000
02/01/2022	US	AMS	110,000
03/01/2022	US	AMS	160,000

My requirement is to extrapolate the Sales data between the 2 given dates based on difference between the 2 dates and concatenate them to the same table.

For example for 2nd Jan, I calculate the date difference between 1st Jan and 1st Feb, which is 31 days and increment Sales amount by (1/31) and for 3rd Jan, I increment Sales amount by (2/31).. so on until 31st Jan.

After the month change, the date difference needs to be recalculated between 1st Feb and 1st March and similar increment on Sales needs to be applied

Answer 1

You can use a UDTF to generate the rows for each day of the month along with an extrapolation of the sales.

create or replace table T1 as 
select 
COLUMN1::date   as Date,
COLUMN2::string as Country,
COLUMN3::string as Region,
COLUMN4::number(38,2) as Sales
from (values
('01/01/2022','UK','EMEA',100000),
('02/01/2022','UK','EMEA',170000),
('03/01/2022','UK','EMEA',120000),
('01/01/2022','US','AMS',90000),
('02/01/2022','US','AMS',110000),
('03/01/2022','US','AMS',160000)
);

create or replace function EXTRAPOLATE_MONTH_TO_DAYS(MONTH_START date, VAL float)
returns table (DAY_DATE date, VAL float)
language javascript
as
$$
{
    initialize: function (argumentInfo, context) {
    },
    processRow: function (row, rowWriter, context) {
        const DAY = 86400000; // Milliseconds per day.
        let firstDay = row.MONTH_START;
        let lastDay = new Date(row.MONTH_START.getFullYear(), row.MONTH_START.getMonth() + 1, 0);
        let daysInMonth = lastDay.getDate();
        let valPerDay = row.VAL / daysInMonth;
        let sum = 0;
        let curDate = row.MONTH_START;
        for (let i = 1; i <= daysInMonth; i++) {
            sum += valPerDay;
            rowWriter.writeRow({DAY_DATE:new Date(firstDay.getTime()+DAY*(i-1)),VAL:sum});
        }
    },
}
$$;


select   DATE
        ,COUNTRY
        ,REGION
        ,SALES
        ,DAY_DATE
        ,round(VAL,2)::number(38,2) EXTRAPOLATED_RUNNING_TOTAL
from T1, table(EXTRAPOLATE_MONTH_TO_DAYS(DATE, SALES::float) over (partition by COUNTRY, REGION order by DATE))
;

First five rows:

DATE	COUNTRY	REGION	SALES	DAY_DATE	EXTRAPOLATED_RUNNING_TOTAL
2022-01-01	US	AMS	90000	2022-01-01	2903.23
2022-01-01	US	AMS	90000	2022-01-02	5806.45
2022-01-01	US	AMS	90000	2022-01-03	8709.68
2022-01-01	US	AMS	90000	2022-01-04	11612.90
2022-01-01	US	AMS	90000	2022-01-05	14516.13