How to add Totals to rows groups generated by a CTE - sql

Question

I have this CTE. My goal is to display all the rows in the result set and the total of a column which has a column with the same value.
WITH CTEVal AS
(SELECT RNum, Amount, Tag, TAmount,
ROW_NUMBER() OVER(ORDER BY CONVERT(DATETIME, TDate, 101) DESC) AS RowNum
from [dbo].[viewJD])
SELECT TDate, RNum,
CASE WHEN Tag = 'DR' THEN CONVERT(VARCHAR(50),Amount) ELSE '0.00' END AS Debit,
CASE WHEN Tag = 'CR' THEN CONVERT(VARCHAR(50),Amount) ELSE '0.00' END AS Credit FROM CTEVal
which returns the result set
TDate Rnum Debit Credit
12-22-2015 8 100 0.00
12-22-2015 8 300 0.00
12-22-2015 8 0.00 400
12-22-2015 9 250 0.00
12-22-2015 9 250 0.00
12-22-2015 9 0.00 500
12-23-2015 10 600 0.00
12-23-2015 10 0.00 100
12-23-2015 10 0.00 500
Expected result set
TDate Rnum Debit Credit
12-22-2015 8 100 0.00
12-22-2015 8 300 0.00
12-22-2015 8 0.00 400
TOTAL 400 400
12-22-2015 9 250 0.00
12-22-2015 9 250 0.00
12-22-2015 9 0.00 500
TOTAL 500 500
12-23-2015 10 600 0.00
12-23-2015 10 0.00 100
12-23-2015 10 0.00 500
TOTAL 600 600
Selecting * from viewJD would return
TDate Rnum Amount Tag TAmount
12-22-2015 8 100 DR 400
12-22-2015 8 300 DR 400
12-22-2015 8 400 CR 400
12-22-2015 9 250 DR 500
12-22-2015 9 250 DR 500
12-22-2015 9 500 CR 500
12-23-2015 10 600 DR 600
12-23-2015 10 100 CR 600
12-23-2015 10 500 CR 600
Sum of DR = sum of CR is = TAmount

Answer 1

I have a solution. I know it is not a fine solution. I have tried using Rollup but could not get the exact result format. So I came up with this.
This is indeed a long solution.
But I hope you would get some idea to solve the problem from this.
Here I have first taken the Debit and Credit separated values into a table variable.
Then I applied SUM to the values and took the complete result to a cte.
I applied DENSE_RANK to get the correct ordering such that the 'total' row comes after the detail rows.
Finally selected records from cte with value 'Total' for column RNum for the row showing 'total'.
CREATE TABLE viewJD
(
TDate Date,
Rnum INT,
Amount INT,
Tag VARCHAR(2),
TAmount INT
)
INSERT INTO viewJD
VALUES
('12-22-2015', 8 , 100 , 'DR', 400),
('12-22-2015' , 8 , 300 , 'DR' , 400),
('12-22-2015', 8 , 400 , 'CR' , 400),
('12-22-2015' , 9 , 250 , 'DR' , 500 ),
('12-22-2015' , 9 , 250 , 'DR' , 500 ),
('12-22-2015' , 9 , 500 , 'CR' , 500 ),
('12-23-2015' , 10 , 600 , 'DR' , 600 ),
('12-23-2015' , 10 , 100 , 'CR' , 600 ),
('12-23-2015' , 10 , 500 , 'CR' , 600 )
DECLARE #ValTable TABLE
(
TDate Date,
Rnum INT,
Debit INT,
Credit INT
)
INSERT INTO #ValTable
SELECT TDate, RNum,
CASE WHEN Tag = 'DR' THEN Amount ELSE 0.00 END AS Debit,
CASE WHEN Tag = 'CR' THEN Amount ELSE 0.00 END AS Credit
FROM viewJD
;WITH cte
AS (
SELECT *,RowOrder=DENSE_RANK() OVER (PARTITION BY NULL ORDER By RNUM) from #ValTable
UNION ALL
SELECT NULL,
RNum ,
SUM([Debit]) AS Debit,
SUM([Credit]) AS Credit ,
RowOrder =DENSE_RANK() OVER (PARTITION BY NULL ORDER By RNUM)
FROM #ValTable
GROUP BY TDate,RNum
)
SELECT TDate,
Rnum = CASE WHEN TDate IS NULL THEN 'Total' ELSE CONVERT(VARCHAR(20),Rnum ) END,
Debit,
Credit
FROM cte
ORDER BY roworder,Rnum,TDate DESC
DROP TABLE viewJD

Answer 2

The format you provide in your expected output is not feasible or advisable. The following code stores them as separate columns.
SELECT TDate,
RNum,
CASE WHEN Tag = 'DR' THEN Amount ELSE 0.00 END AS Debit,
CASE WHEN Tag = 'CR' THEN Amount ELSE 0.00 END AS Credit,
(select sum(Amount)
from [dbo].[viewJD] td2
where td2.RNum = td1.RNum
and Tag = 'DR') as Debit_Total,
(select sum(Amount)
from [dbo].[viewJD] td2
where td2.RNum = td1.RNum
and Tag = 'CR') as Credit_Total
FROM [dbo].[viewJD] td1
This will give you output like
TDate Rnum Debit Credit Debit_Total Credit_Total
12-22-2015 8 100 0.00 400.00 400.00
12-22-2015 8 300 0.00 400.00 400.00
12-22-2015 8 0.00 400 400.00 400.00
12-22-2015 9 250 0.00 500.00 500.00
12-22-2015 9 250 0.00 500.00 500.00
12-22-2015 9 0.00 500 500.00 500.00
12-23-2015 10 600 0.00 600.00 600.00
12-23-2015 10 0.00 100 600.00 600.00
12-23-2015 10 0.00 500 600.00 600.00
Not the nicest to look at, but the easiest to use in other queries.