Need query to subtract the two rows in databricks sql

I have a table and the data is in below format.

I need to derive a 3rd row as mentioned in image. It is a calculated column and first row minus second row. ProgramId, Category, ReportMonth, ReportYear same and third row should hav subcategory 3. Net and it should minus from first row and second row for respective months.

I need the query in databricks sql

I have tried self join and some lag functions but it doesn't work. Please let me know if I am not making any sense.

Thanks, MG

Solution

You can do join on the tables filtered on column Subcategory to get your desired results.

Use below code for it.

SELECT Programid, Category, Subcategory, ReportMon, ReportYea,
       Oct, Nov, Dec, Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep
FROM original_data
UNION ALL
SELECT o.Programid, o.Category, '3. Net' AS Subcategory, '11/2/2023' AS ReportMon, 'FY2024' AS ReportYea,
       o.Oct - e.Oct, o.Nov - e.Nov, o.Dec - e.Dec, o.Jan - e.Jan, o.Feb - e.Feb,
       o.Mar - e.Mar, o.Apr - e.Apr, o.May - e.May, o.Jun - e.Jun, o.Jul - e.Jul, o.Aug - e.Aug, o.Sep - e.Sep
FROM (SELECT * from original_data where Subcategory='1. Adjusted') o
JOIN (SELECT * from original_data where Subcategory='2. Estimated') e
ON o.Programid = e.Programid
AND o.Category = e.Category
AND o.ReportMon = e.ReportMon
AND o.ReportYea = e.ReportYea

Here, i am filtering the records on Subcategory and joining them on same Programid,Category ,ReportMon and ReportYea .

Output:

Programid	Category	Subcategory	ReportMon	ReportYea	Oct	Nov	Dec	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep
12234	ABC	1. Adjusted	11/1/2023	FY2023	23455.55	45464.43	456959.60	154531.60	54521.45	511231.50	35235.67	47575.23	48645.87	54524.89	11353.14	13515.45
12234	ABC	2. Estimated	11/1/2023	FY2023	15311.23	54465.35	53531.45	5453.45	546.40	5856.45	5454.56	452.42	458.50	5661.46	1542.45	5145.67
12234	ABC	3. Net	11/2/2023	FY2024	8144.32	-9000.92	403428.15	149078.15	53975.05	505375.05	29781.11	47122.81	48187.37	48863.43	9810.69	8369.78