October 1, 2012 at 1:34 am
i want to find the increase / decrease amount group by sector and month.
--input table
declare @t table
([id] int,[Amount] int,monthno int,Sector char(1));
INSERT INTO @t
([id],[Amount],Monthno ,sector )
VALUES
(1, 100,1,'A'),
(2,50,1,'A'),
(3, 200,1,'B'),
(4, 300,1,'C'),
(5, 400,2,'A'),
(6, 500,2,'B'),
(7, 600,2,'C'),
(8, 100,3,'A'),
(9, 500,3,'B'),
(10, 100,3,'C')
select * from @t
-- Required Output
declare @output table ([IncreaseDecreaseAmount] int,Amount int,[Monthno]int,sector char(1))
INSERT INTO @output
([IncreaseDecreaseAmount],Amount,[Monthno],sector )
VALUES
(50,100,1,'A'),
(300,400,2,'A'),
(-200,100,3,'A'),
(200,200,1,'B'),
(300,500,2,'B'),
(0,500,3,'B')
select * from @output
regards,
October 1, 2012 at 2:52 am
What are the rules for calculating increase & decrease?
It's not obvious from the expected results.
October 1, 2012 at 5:27 am
the rule here is we have month no here.
1. check secondmonth amount
2. sum(second month amount) > sum(first month amount)
3. get the difference
4. difference is greater then sum(first month amount) then it shows addition.
5. difference is less then sum(first month amount) then it shows reduction
6. check the amount for month3
7. sum of (First Month+second month) > sum of (third month)
8. difference is greater then sum(first+second month amount) then it shows addition.
9. difference is less then sum(first+second month amount) then it shows reduction
it shall be grouped by monthno and sector.
October 2, 2012 at 2:54 am
sayedkhalid99 (10/1/2012)
the rule here is we have month no here.1. check secondmonth amount
2. sum(second month amount) > sum(first month amount)
3. get the difference
4. difference is greater then sum(first month amount) then it shows addition.
5. difference is less then sum(first month amount) then it shows reduction
6. check the amount for month3
7. sum of (First Month+second month) > sum of (third month)
8. difference is greater then sum(first+second month amount) then it shows addition.
9. difference is less then sum(first+second month amount) then it shows reduction
it shall be grouped by monthno and sector.
I still can't follow that.
Can you give the values for Sector A with the steps above? That should make it clearer.
October 2, 2012 at 4:51 am
-- you have 2 rows for sector 'A', month 1
-- what do you want to do with them?
-- It's not clear from your "Required Output", which is incorrect
-- see row 3.
SELECT
[IncreaseDecreaseAmount] = ISNULL(cr.Amount-lr.Amount,cr.Amount),
cr.Amount,
cr.Monthno,
cr.sector
FROM @t cr
LEFT JOIN (
SELECT Sector, monthno, [Amount] = SUM([Amount])
FROM @t
GROUP BY Sector, monthno
) lr ON lr.Sector = cr.Sector AND lr.monthno+1 = cr.monthno
ORDER BY cr.sector, cr.Monthno
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
October 2, 2012 at 10:53 pm
i have changed the my input and output and add date and comments to make it clear, for multiple entry with in a month it shall take the last entry.
declare @t table
([id] int,[Amount] int,Sector char(1),Currentdate datetime);
INSERT INTO @t
([id],[Amount],Sector,currentdate )
VALUES
(1, 100,'A','2002-01-01 00:00:00.000'),
(2,50,'A','2002-01-02 00:00:00.000'),
(3, 200,'B','2002-01-01 00:00:00.000'),
(4, 300,'C','2002-01-01 00:00:00.000'),
(5, 400,'A','2002-02-02 00:00:00.000'),
(6, 500,'B','2002-02-02 00:00:00.000'),
(7, 600,'C','2002-02-02 00:00:00.000'),
(8, 100,'A','2002-03-02 00:00:00.000'),
(9, 500,'B','2002-03-03 00:00:00.000'),
(10, 100,'C','2002-03-03 00:00:00.000'),
(11, 200,'C','2002-03-03 00:00:00.000')
select * from @t
declare @output table ([AfterIncreaseDecreaseAmount] int,Amount int,[Monthno]int,sector char(1))
INSERT INTO @output
([AfterIncreaseDecreaseAmount],Amount,[Monthno],sector )
VALUES
(50,100,1,'A'), -- for 01/01/2002 it is 100 , for 02/02/2002 for multiple entry with in the month it shall take the last amount based on max(id) or last(currentdate).
(+350,400,2,'A'), -- for second month 02/02/2002 it 400 > 50 it shows increase from first month so 400-50=+350 so --> AfterIncreaseDecreaseAmount=350
(-300,100,3,'A'), -- for third month 02/03/2002 it is 100 < 400 (sum of 1st,2st month amount) so it shows decrease 100-400=-300 so-->AfterIncreaseDecreaseAmount=-300
(+200,200,1,'B'),-- for b we have 200 it is 200 > 0 ,if it is only one it shall count that as 0 so it shows increase 200-0=200 --> AfterIncreaseDecrease=200
(+300,500,2,'B'),-- for b we have 500 500 > 200--> 500-200=300 it show increased -->AfterIncreaseDecrease=300
(0,500,3,'B'), -- for b in third month 500-500(sum of 1st,2stmonth amount of b) =0 AfterIncreaseDecrease=0
(+300,300,1,'C'),
(+300,600,2,'C'),
(400,100,3,'C')
select * from @output
October 3, 2012 at 7:39 am
sayedkhalid99 (10/2/2012)
i have changed the my input and output and add date and comments to make it clear, for multiple entry with in a month it shall take the last entry.....
Please confirm that your desired results are correct for sector C, and add a few extra sample rows for B and C, including multiple rows for a single month. It will aid disambiguation. Thanks.
EDIT: more sample data requested.
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply