distribution of income over time based on rule set - sql-server

Question

Not sure if I was able to properly describe the title of my question but I haven't done anything this complex in T-Sql. Here's what I'm trying to achieve.
After several queries and table building, I got the data (below) in a temp table which probably will become a permanent table.
The part I need help with is, Need to create another table to report part of the sale from the data below.
Table 1
LOC# ITEM sale_total FiscalYear FiscalPrd TYPE O_R DESC LFT_EXP METHOD PRD_MO AG_PCNT PLAN_ID YEARS STATUS
33 17879 12203.57 2017 3 A O Plan A L W 120 100.00 6 10 N
33 17903 579.97 2017 3 A O Plan A L W 120 100.00 6 10 N
33 18870 799.60 2017 3 A O Plan A L W 120 100.00 6 10 N
33 19739 319.95 2017 3 A B Plan C E F 36 100.00 8 3 N
33 21585 559.96 2017 3 B O Plan B E N 0 100.00 10 0 N
A percentage of the sale_total from table above needs to be distributed based on the table below (I can also include the YR01-YR10 columns in the table above if it make things simpler).
Table 2
PLAN_ID EFDATE MONTHS METHOD EPERCENT YR01 YR02 YR03 YR04 YR05 YR06 YR07 YR08 YR09 YR10
6 2007-08-01 120 W 100.00 40.00 20.00 11.00 8.00 6.00 5.00 4.00 3.00 2.00 1.00
8 2012-11-12 36 F 100.00 33.00 33.00 34.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
10 2015-10-01 0 N 100.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
So my end result would look like below for LOC#33 and item 18870: (So the calculation is based on the plan. For this specific example, item 18870 falls under plan 6 (table 1). Table 2 has the distribution of all plans. We can see it spreads over 120 months and YR01-YR10 has the yearly percent distribution. Now from the sales number for 18870 from table 1, 799.60, distribution math for year 1 is (799.60 x 0.4) / 12 = 26.65. Same method follows up to year 10 for this plan and up to year 3 for plan 8. For the 10 year plan, it will create 120 input records. 1 for each month for each item, location)
LOC# ITEM FiscalYear FiscalPrd SALES
33 18870 2017 4 26.65
33 18870 2017 5 26.85
33 18870 2017 6 26.85
.............................................
.............................................
33 18870 2017 11 26.85
33 18870 2017 12 26.85
33 18870 2018 1 13.33
33 18870 2018 2 13.33
.............................................
.............................................
33 18870 2018 11 13.33
33 18870 2018 12 13.33
.............................................
.............................................
33 18870 2027 2 0.67
33 18870 2027 3 0.67
It will distribute that income over 120 months at the percentage dictated by the table 2 for planID 6. for planID 8, it will do the same math over 36 months.
The percentages can change over time and a user will manage the plans. There are other locations and the first table will include all location for the current fiscal period and all items (3 max as shows in Table 2). I created a status columns so I can mark if once the data is processed.
When comes the next fiscal period sales, the program will update the above result set sales numbers to include new sales. It will be run once a month to generate the above result.
Would greatly appreciate some direction as I got up to this point and not sure how to approach from here.
Thanks a bunch.
#ZLK I saved the first run in a temp table, then added the following rows
(33, 17879, 10000.00, 2017, 4, 6, 10),
(33, 17903, 500.00, 2017, 4, 6, 10),
(33, 18870, 800.00, 2017, 4, 6, 10),
(33, 19739, 300.00, 2017, 4, 8, 3),
(33, 21585, 500.00, 2017, 4, 10, 0);
This the save first set of item, location with different sales number in fiscal period 4. the first set was inserted into a table #finalDist, the second uses the following query to insert and update.
begin tran
if exists (select * FROM #finalDist AGE
INNER JOIN #dist PrdAg on PrdAg.Loc = AGE.Loc
and PrdAg.Item = AGE.Item
and PrdAg.FiscalYear = AGE.FiscalYear
and PrdAg.FiscalPrd = AGE.FiscalPrd)
begin
UPDATE AGE
SET AGE.SALES = AGE.SALES + PrdAg.SALES
FROM #finalDist AGE
INNER JOIN #dist PrdAg on PrdAg.Loc = AGE.Loc
and PrdAg.Item = AGE.Item
and PrdAg.FiscalYear = AGE.FiscalYear
and PrdAg.FiscalPrd = AGE.FiscalPrd
end
else
begin
INSERT INTO #finalDist (Loc, Item, FiscalYear, FiscalPrd, SALES)
SELECT Loc, Item, FiscalYear, FiscalPrd, SALES from #dist PrdAg
WHERE NOT EXISTS (SELECT AGE.Item FROM #finalDist AGE WHERE PrdAg.Loc = AGE.Loc
and PrdAg.Item = AGE.Item
and PrdAg.FiscalYear = AGE.FISCALYear
and PrdAg.FiscalPrd = AGE.FiscalPrd)
end
commit tran
I realize I can't do if else as I need to do both insert and update. It will update everything except the fiscal period 3, then it will insert four rows for this case
Loc Item FiscalYear FiscalPrd SALES
33 18870 2027 4 0.66666666666
33 19739 2020 4 8.50000000000
33 17903 2027 4 0.41666666666
33 17879 2027 4 8.33333333333
update look like this
33 17879 2017 4 406.78566666666 -- this row wasn't updated
33 17879 2017 5 740.11899999999

Answer 1

One way you can do it is like this:
DECLARE #T TABLE (Loc INT, Item INT, Sales_Total DECIMAL(10, 2), FiscalYear INT, FiscalPrd INT, Plan_ID INT, YEARS INT);
INSERT #T (Loc, Item, Sales_Total, FiscalYear, FiscalPrd, Plan_ID, YEARS) VALUES
(33, 17879, 12203.57, 2017, 3, 6, 10),
(33, 17903, 579.97, 2017, 3, 6, 10),
(33, 18870, 799.60, 2017, 3, 6, 10),
(33, 19739, 319.95, 2017, 3, 8, 3),
(33, 21585, 559.96, 2017, 3, 10, 0);
DECLARE #T2 TABLE (PLAN_ID INT, YR01 DECIMAL(4, 2), YR02 DECIMAL(4, 2), YR03 DECIMAL(4, 2), YR04 DECIMAL(4, 2), YR05 DECIMAL(4, 2), YR06 DECIMAL(4, 2), YR07 DECIMAL(4, 2), YR08 DECIMAL(4, 2), YR09 DECIMAL(4, 2), YR10 DECIMAL(4, 2));
INSERT #T2 (PLAN_ID, YR01, YR02, YR03, YR04, YR05, YR06, YR07, YR08, YR09, YR10) VALUES
(6, 40.00, 20.00, 11.00, 8.00, 6.00, 5.00, 4.00, 3.00, 2.00, 1.00),
(8, 33.00, 33.00, 34.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00),
(10, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00);
WITH CTE(N) AS (SELECT 1 FROM (VALUES (1), (1), (1), (1), (1), (1), (1), (1), (1), (1), (1)) AS A(B)),
CTE2(N) AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) FROM CTE AS A CROSS JOIN CTE) -- Numbers 1-121
SELECT T.Loc,
T.Item,
FiscalYear = DATEPART(YEAR, DATEADD(MONTH, CTE2.N, DATEFROMPARTS(T.FiscalYear, T.FiscalPrd, 1))),
FiscalPrd = DATEPART(MONTH, DATEADD(MONTH, CTE2.N, DATEFROMPARTS(T.FiscalYear, T.FiscalPrd, 1))),
SALES = T.Sales_Total * C.Vals / 100 / 12
FROM #T AS T
JOIN CTE2
ON CTE2.N <= T.YEARS * 12
JOIN
(
SELECT T.PLAN_ID, C.Y, C.Vals
FROM #T2 AS T
CROSS APPLY (VALUES (1, YR01), (2, YR02), (3, YR03), (4, YR04), (5, YR05), (6, YR06), (7, YR07), (8, YR08), (9, YR09), (10, YR10)) AS C(Y, Vals) -- Unpivot
) AS C
ON C.PLAN_ID = T.PLAN_ID
AND C.Y = (CTE2.N + 11) / 12
ORDER BY T.Loc, T.Item, CTE2.N;
Essentially you want to get a list of numbers (up to 120 here but if there's more than 10 years, you would need more) and cross join it to your table. Then unpivot the values in table2 and link them to the numbers table in (grouping them by years).
EDIT:
Taking into account that values can change and assuming the end date is still the same and doesn't increase, this is what I believe you're looking for:
DECLARE #T TABLE (Loc INT, Item INT, Sales_Total DECIMAL(10, 2), FiscalYear INT, FiscalPrd INT, Plan_ID INT, YEARS INT);
INSERT #T (Loc, Item, Sales_Total, FiscalYear, FiscalPrd, Plan_ID, YEARS) VALUES
(33, 17879, 12203.57, 2017, 3, 6, 10),
(33, 17903, 579.97, 2017, 3, 6, 10),
(33, 18870, 799.60, 2017, 3, 6, 10),
(33, 19739, 319.95, 2017, 3, 8, 3),
(33, 21585, 559.96, 2017, 3, 10, 0),
(33, 17879, 10000.00, 2017, 4, 6, 10),
(33, 17903, 500.00, 2017, 4, 6, 10),
(33, 21585, 500.00, 2017, 4, 10, 0),
(33, 17879, 5000.00, 2017, 5, 6, 10),
(33, 17879, 6900.00, 2017, 7, 6, 10);
DECLARE #T2 TABLE (PLAN_ID INT, YR01 DECIMAL(4, 2), YR02 DECIMAL(4, 2), YR03 DECIMAL(4, 2), YR04 DECIMAL(4, 2), YR05 DECIMAL(4, 2), YR06 DECIMAL(4, 2), YR07 DECIMAL(4, 2), YR08 DECIMAL(4, 2), YR09 DECIMAL(4, 2), YR10 DECIMAL(4, 2));
INSERT #T2 (PLAN_ID, YR01, YR02, YR03, YR04, YR05, YR06, YR07, YR08, YR09, YR10) VALUES
(6, 40.00, 20.00, 11.00, 8.00, 6.00, 5.00, 4.00, 3.00, 2.00, 1.00),
(8, 33.00, 33.00, 34.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00),
(10, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00);
WITH CTE(N) AS (SELECT 1 FROM (VALUES (1), (1), (1), (1), (1), (1), (1), (1), (1), (1), (1)) AS A(B)),
CTE2(N) AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) - 1 FROM CTE AS A CROSS JOIN CTE) -- Numbers 0-120
SELECT T.Loc,
T.Item,
FiscalYear = DATEPART(YEAR, DATEADD(MONTH, CTE2.N, T.FirstDate)),
FiscalPrd = DATEPART(MONTH, DATEADD(MONTH, CTE2.N, T.FirstDate)),
SALES = T.TotalToDate * C.Vals / 100 / 12
FROM
(
SELECT T.Loc,
T.Item,
T.YEARS,
T.Plan_ID,
C.TotalToDate,
FirstDate = CAST(DATEADD(MONTH, 1, CAST(MIN(FiscalYear) OVER (PARTITION BY Loc, Item) AS CHAR(4)) + '-' +
CAST(MIN(FiscalPrd) OVER (PARTITION BY Loc, Item) AS VARCHAR(2)) + '-01') AS DATE),
CurrentDate = CAST(DATEADD(MONTH, 1, CAST(FiscalYear AS CHAR(4)) + '-' + CAST(FiscalPrd AS VARCHAR(2)) + '-01') AS DATE),
C2.NextDate
FROM #T AS T
CROSS APPLY
(
SELECT SUM(Sales_Total)
FROM #T
WHERE Loc = T.Loc
AND Item = T.Item
AND FiscalYear * 12 + FiscalPrd <= T.FiscalYear * 12 + T.FiscalPrd
) AS C(TotalToDate)
OUTER APPLY
(
SELECT CAST(DATEADD(MONTH, 1, CAST(MIN(FiscalYear) AS CHAR(4)) + '-' +
CAST(MIN(FiscalPrd) AS VARCHAR(2)) + '-01') AS DATE)
FROM #T
WHERE Loc = T.Loc
AND Item = T.Item
AND FiscalYear * 12 + FiscalPrd > T.FiscalYear * 12 + T.FiscalPrd
) AS C2(NextDate)
) AS T
JOIN CTE2
ON CTE2.N <= T.YEARS * 12 - 1
AND CTE2.N >= DATEDIFF(MONTH, T.FirstDate, T.CurrentDate)
AND DATEADD(MONTH, CTE2.N, T.FirstDate) < COALESCE(T.NextDate, DATEADD(MONTH, T.YEARS * 12, T.FirstDate))
JOIN
(
SELECT T.PLAN_ID, C.Y, C.Vals
FROM #T2 AS T
CROSS APPLY (VALUES (1, YR01), (2, YR02), (3, YR03), (4, YR04), (5, YR05), (6, YR06), (7, YR07), (8, YR08), (9, YR09), (10, YR10)) AS C(Y, Vals) -- Unpivot
) AS C
ON C.PLAN_ID = T.PLAN_ID
AND C.Y = (CTE2.N + 12) / 12
ORDER BY T.Loc, T.Item, CTE2.N;
SECOND EDIT (Modified to have the start date be the month declared, e.g. March 2017 means the plan starts in March 2017 instead of April 2017. The only difference is adjusting the row_number joins):
/*
#PlanTable: A static table for plans.
*/
DECLARE #PlanTable TABLE (PLAN_ID INT, YR01 DECIMAL(4, 2), YR02 DECIMAL(4, 2), YR03 DECIMAL(4, 2), YR04 DECIMAL(4, 2), YR05 DECIMAL(4, 2), YR06 DECIMAL(4, 2), YR07 DECIMAL(4, 2), YR08 DECIMAL(4, 2), YR09 DECIMAL(4, 2), YR10 DECIMAL(4, 2));
INSERT #PlanTable (PLAN_ID, YR01, YR02, YR03, YR04, YR05, YR06, YR07, YR08, YR09, YR10) VALUES
(6, 40.00, 20.00, 11.00, 8.00, 6.00, 5.00, 4.00, 3.00, 2.00, 1.00),
(8, 33.00, 33.00, 34.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00),
(10, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00);
/*
#RecordsToInsert: The values you're inserting into your final table. Starting with FiscalYear 2017 and FiscalPrd 3 as an example.
*/
DECLARE #RecordsToInsert TABLE (Loc INT, Item INT, Sales_Total DECIMAL(10, 2), FiscalYear INT, FiscalPrd INT, Plan_ID INT, YEARS INT);
INSERT #RecordsToInsert (Loc, Item, Sales_Total, FiscalYear, FiscalPrd, Plan_ID, YEARS) VALUES
(33, 17879, 12203.57, 2017, 3, 6, 10),
(33, 17903, 579.97, 2017, 3, 6, 10),
(33, 18870, 799.60, 2017, 3, 6, 10),
(33, 19739, 319.95, 2017, 3, 8, 3),
(33, 21585, 559.96, 2017, 3, 10, 0);
/*
#FinalTable: This is the table you're using to insert the data.
*/
DECLARE #FinalTable TABLE (Loc INT, Item INT, FiscalYear INT, FiscalPrd INT, SALES DECIMAL(10, 2));
/*
Start by inserting the first lot of information into the final table.
*/
WITH CTE(N) AS (SELECT 1 FROM (VALUES (1), (1), (1), (1), (1), (1), (1), (1), (1), (1), (1)) AS A(B)),
CTE2(N) AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) - 1 FROM CTE AS A CROSS JOIN CTE) -- Numbers 0-120
INSERT #FinalTable (Loc, Item, FiscalYear, FiscalPrd, SALES)
SELECT T.Loc,
T.Item,
FiscalYear = DATEPART(YEAR, DATEADD(MONTH, CTE2.N, CAST(T.FiscalYear AS CHAR(4)) + '-' + CAST(T.FiscalPrd AS VARCHAR(2)) + '-01')),
FiscalPrd = DATEPART(MONTH, DATEADD(MONTH, CTE2.N, CAST(T.FiscalYear AS CHAR(4)) + '-' + CAST(T.FiscalPrd AS VARCHAR(2)) + '-01')),
SALES = T.Sales_Total * C.Vals / 100 / 12
FROM #RecordsToInsert AS T
JOIN CTE2
ON CTE2.N <= T.YEARS * 12 - 1
JOIN
(
SELECT T.PLAN_ID, C.Y, C.Vals
FROM #PlanTable AS T
CROSS APPLY (VALUES (1, YR01), (2, YR02), (3, YR03), (4, YR04), (5, YR05), (6, YR06), (7, YR07), (8, YR08), (9, YR09), (10, YR10)) AS C(Y, Vals) -- Unpivot
) AS C
ON C.PLAN_ID = T.PLAN_ID
AND C.Y = CTE2.N / 12 + 1
ORDER BY T.Loc, T.Item, CTE2.N;
/*
Next we want to insert records for the next FiscalPrd, so we get rid of the last ones and put in FiscalYear 2017, FiscalPrd 4.
*/
DELETE #RecordsToInsert;
INSERT #RecordsToInsert (Loc, Item, Sales_Total, FiscalYear, FiscalPrd, Plan_ID, YEARS) VALUES
(33, 17879, 10000.00, 2017, 4, 6, 10),
(33, 17903, 500.00, 2017, 4, 6, 10),
(33, 18870, 800.00, 2017, 4, 6, 10),
(33, 19739, 300.00, 2017, 4, 8, 3),
(33, 21585, 500.00, 2017, 4, 10, 0);
/*
Then we update the final table with the new sales total, ignoring periods that were prior to the fiscal year and period being inserted.
*/
UPDATE T
SET SALES = SALES_NEW
FROM (
SELECT FT.Loc,
FT.Item,
FT.FiscalYear,
FT.FiscalPrd,
FT.SALES,
SALES_NEW = (SUM(FT.SALES) OVER (PARTITION BY FT.Loc, FT.Item) + R.Sales_Total) * C.Vals / 100 / COUNT(1) OVER (PARTITION BY FT.Loc, FT.Item, (FT.RN + 11) / 12)
FROM (SELECT *, RN = ROW_NUMBER() OVER (PARTITION BY Loc, Item ORDER BY FiscalYear, FiscalPrd) FROM #FinalTable) AS FT
JOIN #RecordsToInsert AS R
ON R.Loc = FT.Loc
AND R.Item = FT.Item
AND R.FiscalYear * 12 + R.FiscalPrd <= FT.FiscalYear * 12 + FT.FiscalPrd
JOIN
(
SELECT T.PLAN_ID, C.Y, C.Vals
FROM #PlanTable AS T
CROSS APPLY (VALUES (1, YR01), (2, YR02), (3, YR03), (4, YR04), (5, YR05), (6, YR06), (7, YR07), (8, YR08), (9, YR09), (10, YR10)) AS C(Y, Vals) -- Unpivot
) AS C
ON C.PLAN_ID = R.PLAN_ID
AND C.Y = (FT.RN + 11) / 12
) T;
/*
For the first year, the sales for 2017/5 onwards have changed to (totalsales - firstmonth) * 40% / 11
For the second years onward, the sales have changed to totalsales * whatever% / 12
The first month of the first year remains the same.
*/
SELECT Loc, Item, FiscalYear, FiscalPrd, SALES
FROM #FinalTable
ORDER BY Loc, Item, FiscalYear, FiscalPrd;