Cumulative total for quarters - sql

Question

I currently have a query which produces data for each quarter, what I would like is for the total to be a running total. So quarter 2's figures will be added on to quarter 1's and so on until the end of the fiscal year, where the process will then start again.
select datepart(year,date) as Year,
CASE
WHEN MONTH(date) BETWEEN 1 AND 3 THEN convert(char(4), YEAR(date) - 1) + 'Q4'
WHEN MONTH(date) BETWEEN 4 AND 6 THEN convert(char(4), YEAR(date) - 0) + 'Q1'
WHEN MONTH(date) BETWEEN 7 AND 9 THEN convert(char(4), YEAR(date) - 0) + 'Q2'
WHEN MONTH(date) BETWEEN 10 AND 12 THEN convert(char(4), YEAR(date) - 0) + 'Q3'
END AS Quarter,
Data1,
Data2,
FROM TABLE
GROUP BY datepart(year,date),
CASE
WHEN MONTH(date) BETWEEN 1 AND 3 THEN convert(char(4), YEAR(date) - 1) + 'Q4'
WHEN MONTH(date) BETWEEN 4 AND 6 THEN convert(char(4), YEAR(date) - 0) + 'Q1'
WHEN MONTH(date) BETWEEN 7 AND 9 THEN convert(char(4), YEAR(date) - 0) + 'Q2'
WHEN MONTH(date) BETWEEN 10 AND 12 THEN convert(char(4), YEAR(date) - 0) + 'Q3'
END
This is what I currently get
year quarter data1 data2
2016 Q1 10 4
2016 Q2 5 6
2016 Q3 4 2
2017 Q4 1 1
I would like the output to look like this
year quarter data1 data2
2016 Q1 10 4
2016 Q2 15 10
2016 Q3 19 12
2017 Q4 20 13
Thanks for any help

Answer 1

You can wrap your query in a CTE and use CROSS APPLY in order to calculate the cumulative sum:
;WITH CTE AS (
-- your query
)
SELECT t1.year, t1.quarter, t1.data1, t3.cumulative_sum
FROM CTE AS t1
CROSS APPLY (
SELECT SUM(data2) AS cumulative_sum
FROM CTE AS t2
WHERE t2.year <= t1.year AND t2.quarter <= t1.quarter) AS t3

Answer 2

First, your query can be simplified to:
SELECT datepart(year, date) as Year,
datename(year, date) + 'Q' + datename(quarter, dateadd(month, -3, date)) as quarter,
SUM(Data1) as Data1, SUM(Data2) as Data2
FROM TABLE
GROUP BY datepart(year, date) as Year,
datename(year, date) + 'Q' + datename(quarter, dateadd(month, -3, date)) as quarter
ORDER BY quarter;
Note the use of datename(). This returns a string rather than a number, which is easier for string concatenation. In addition, this simplifies the logic just by subtracting 3 months to get the quarter.
In SQL Server 2012+, you can use the built in functionality for cumulative sum. In SQL Server 2008, one method uses apply:
WITH t as (
SELECT datepart(year, date) as Year,
datename(year, date) + 'Q' + datename(quarter, dateadd(month, -3, date)) as quarter,
SUM(Data1) as Data1, SUM(Data2) as Data2
FROM TABLE
GROUP BY datepart(year, date) as Year,
datename(year, date) + 'Q' + datename(quarter, dateadd(month, -3, date)) as quarter
)
SELECT t.*, tt.data2
FROM t OUTER APPLY
(SELECT SUM(data1) as data2
FROM t t2
WHERE t2.quarter <= t.quarter
) tt;