Sum or join rows by criteria - sql

Question

Suppose I have a table like this:
Table1
EmployeeID Month sym Quantity V_M BudjetCode
222222 1 40 133.35 1214 800000
222222 2 40 178.50 115 800000
222222 3 40 150.33 215 800000
222222 4 40 186.37 315 800000
222222 5 40 127.38 415 800000
222222 6 40 153.00 515 800000
222222 7 40 178.50 615 800000
222222 7 40 8.50 615 700052015
222222 8 40 187.00 715 800000
And I would like to change this table to:
Table 2
EmployeeID Month sym Quantity V_M BudjetCode
222222 1 40 133.35 1214 800000
222222 2 40 178.50 115 800000
222222 3 40 150.33 215 800000
222222 4 40 186.37 315 800000
222222 5 40 127.38 415 800000
222222 6 40 161.5 515 800000
222222 7 40 178.50 615 800000
222222 8 40 187.00 715 800000
How?
See the row in table 1 where the BudjetCode is unusual?
well, for this row I would like to add the Quantity value (8.5) to the row there V_M is one less than the V_M is the original row (where the budjetCode is 700052015).
In the example, in the original it was 615 so one less is 515 (615 means date 6.15 and 515 means 5.15) and to that I need to add the quantity (8.5 to 153 = 161.5)
I was thinking of "over partition":
select [EmployeeID],[Month],[Sym],
sum([Quantity])
over (partition by [EmployeeID], [V_M]-1 where???) as b
from table1
where [Sym] = '40' and [EmployeeID] = 222222
order by [Month]
But I don't know how to sum this up usiing the criteria of the budjetCode starts with "700".
comment : Don't to this for the first row.
Update:
EmployeeID month Quantity V_M MS_BudjetCode
22222 1 40 133.35 1214 88888888
22222 2 40 178.50 115 88888888
22222 3 40 150.33 215 88888888
22222 4 40 186.37 315 88888888
22222 5 40 127.38 415 88888888
22222 6 40 153.00 515 88888888
22222 7 40 8.50 615 700000000
22222 7 40 178.50 615 88888888
22222 8 40 187.00 715 88888888
Output:
22222 2 40 178.50 115 88888888
22222 4 40 186.37 315 88888888
22222 8 40 187.00 715 88888888
22222 3 40 151.33 215 88888888
22222 3 40 151.33 215 88888888
22222 3 40 151.33 215 88888888
22222 3 40 149.33 215 88888888
22222 3 40 149.33 215 88888888
22222 5 40 127.38 415 88888888
22222 6 40 152.00 515 88888888
22222 6 40 154.00 515 88888888
22222 6 40 152.00 515 88888888
22222 6 40 154.00 515 88888888
22222 6 40 154.00 515 88888888
22222 6 40 152.00 515 88888888
22222 6 40 161.50 515 88888888
22222 7 40 178.50 615 88888888
22222 1 40 133.35 1214 88888888
Query:
SELECT t1.EmployeeID, t1.Month, t1.sym,
t1.Quantity + COALESCE(t2.Quantity, 0),
t1.V_M, t1.BudjetCode
FROM Table1 AS t1
LEFT JOIN Table1 AS t2
ON t1.EmployeeID = t2.EmployeeID AND t1.V_M = t2.V_M - 100 AND
SUBSTRING(t2.BudjetCode,1,3) = '700'
WHERE SUBSTRING(t1.BudjetCode,1,3) <> '700' and sym='40' and EmployeeID = 22222

Answer 1

This may be a solution, i think we cant simply subtract 1 with month, need to consider january december cases also. i had included that checking also. here budgetcode is distinguished with length, you can change it if needed.
;WITH CTETable1 AS
(
SELECT
EmployeeID,[Month],sym,
CASE WHEN LEN(BudjetCode)>6 THEN Quantity ELSE 0 END Quantity,
CASE WHEN LEN(BudjetCode)>6 THEN
(CASE WHEN LEFT(V_M,LEN(V_M)-2)=1 THEN 12 ELSE LEFT(V_M,LEN(V_M)-2)-1 END)*100+
(CASE WHEN LEFT(V_M,LEN(V_M)-2)=1 THEN RIGHT(V_M,2)-1 ELSE RIGHT(V_M,2) END) ELSE 0 END V_M,
BudjetCode
FROM Table1
)
SELECT
T1.EmployeeID,T1.[Month],T1.sym,T1.Quantity+ISNULL(T2.Quantity,0) Quantity,
T1.V_M,T1.BudjetCode
FROM Table1 T1
LEFT JOIN CTETable1 T2 ON T1.EmployeeID=T2.EmployeeID AND T1.V_M=T2.V_M
WHERE LEN(T1.BudjetCode)=6

Answer 2

Will this solve your problem?
DELETE
FROM
table
WHERE BudjetCode='700052015'
AND
UPDATE table
SET Quantity = (Quantity+8.5)
WHERE Month = 6;

Answer 3

I think you can do what you want using a simple LEFT JOIN operation:
SELECT t1.EmployeeID, t1.Month, t1.sym,
t1.Quantity + COALESCE(t2.Quantity, 0),
t1.V_M, t1.BudjetCode
FROM Table1 AS t1
LEFT JOIN Table1 AS t2
ON t1.EmployeeID = t2.EmployeeID AND t1.V_M = t2.V_M - 100 AND
SUBSTRING(t2.BudjetCode,1,3) = '700'
WHERE SUBSTRING(t1.BudjetCode,1,3) <> '700'
The query select all table rows but the 'unusual' ones and joins them with 'unusual' rows one day (or is it maybe hour?) ahead. Quantity returned is the sum of the value of 'normal' row plus the value of the related 'unusual' row (if such one really exists).
Expression:
t1.V_M = t2.V_M - 100
implements one less relation between V_Mof t1 and t2. You may change it to suit your actual needs.
Demo here