Average payment delay - year to date - sql

Question

I need to create a query that will allow me to determine the average payment delay per customer and I will hate to update that query each week.
Therefore I already calculated the delay with the date of the document and the date of the payment.
What I need to do know is to have a graph that will show the average of these delays per week but taking into account the previous weeks:
Week 1: average of delays from week 1
Week 2: average of delays from week 1 to week 2
Week 3: average of delays from week 1 to week 3
and so on.
Today, we are week 11 so next week, I will need to have the result of week 12 automatically in my graph.
I already tried a graph using "Running Total in" in the settings of the pivot table
but this result shows on:
Which is not helping because it's making a sum of the delays...
Here is how my code looks like:
SELECT ch.HDOCNO,
ch.HDOCDATE,
ch.HYEAR,
week(ch.HDOCDATE)-1 as "Week",
ch.HMDATE as "Payment date",
AVG(ch.HMDATE-ch.HDOCDATE) as "Delay"
from AC_CHISTO ch
where ch.HFYEAR = '2016'
and ch.HMDATE IS NOT NULL
and UPPER(ch.HDBK) = 'VEN'
GROUP BY ch.HDOCNO, ch.HDOCDATE, ch.HYEAR, ch.HMDATE
Here is an example of the data that I need to use:

Answer 1

The "date" table that Steve mentions is a fairly straightforward way to solve this problem. The basic premise is to use a pre-computed table to store all possible weeks, making it easy for you to aggregate your results over valid weeks. I've included comments in the code below to identify the different parts that I had to include just to return some example results using the data you provided.
--Create table to store source data (not part of a final solution, only needed in this case to return example data).
DECLARE #AC_CHISTO TABLE
(
HDOCNO int,
HDOCDATE datetime,
HYEAR varchar(4),
HMDATE datetime,
HDBK varchar(3)
)
INSERT INTO #AC_CHISTO
SELECT 610474, '02/26/2016', '2016', '03/02/2016', 'VEN'
UNION ALL
SELECT 611727, '03/04/2016', '2016', '03/11/2016', 'VEN'
UNION ALL
SELECT 611728, '03/04/2016', '2016', '03/09/2016', 'VEN'
UNION ALL
SELECT 6133119, '03/11/2016', '2016', '03/15/2016', 'VEN'
UNION ALL
SELECT 613120, '03/11/2016', '2016', '03/15/2016', 'VEN'
UNION ALL
SELECT 601019, '01/07/2016', '2016', '01/29/2016', 'VEN'
UNION ALL
SELECT 603591, '01/21/2016', '2016', '02/29/2016', 'VEN'
UNION ALL
SELECT 600195, '01/04/2016', '2016', '01/21/2016', 'VEN'
UNION ALL
SELECT 600732, '01/06/2016', '2016', '01/21/2016', 'VEN'
UNION ALL
SELECT 601921, '01/13/2016', '2016', '01/28/2016', 'VEN'
UNION ALL
SELECT 602561, '01/18/2016', '2016', '01/28/2016', 'VEN'
UNION ALL
SELECT 603451, '01/21/2016', '2016', '02/11/2016', 'VEN'
--A table containing all weeks. In the final solution, this should be a real table prepopulated with 1 through 52, rather than a table variable.
DECLARE #AllWeeks TABLE
(
[Week] int
)
INSERT INTO #AllWeeks
SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4 UNION ALL SELECT 5 UNION ALL SELECT 6 UNION ALL SELECT 7 UNION ALL SELECT 8 UNION ALL SELECT 9 UNION ALL SELECT 10 UNION ALL SELECT 11 UNION ALL SELECT 12 UNION ALL SELECT 13 UNION ALL SELECT 14 UNION ALL SELECT 15
--The actual solution begins here.
DECLARE #MaxWeek int
SELECT #MaxWeek = MAX(DATEPART(WK, ch.HDOCDATE)-1) FROM #AC_CHISTO ch
SELECT
aw.[Week],
AVG(results.[Delay]) AS AvgDelay
FROM #AllWeeks aw
INNER JOIN
(
SELECT ch.HDOCNO,
ch.HDOCDATE,
ch.HYEAR,
DATEPART(WK, ch.HDOCDATE)-1 as "Week",
ch.HMDATE as "PaymentDate",
DATEDIFF(DD, ch.HDOCDATE, ch.HMDATE) as "Delay"
FROM #AC_CHISTO ch
WHERE ch.HYEAR = '2016'
and ch.HMDATE IS NOT NULL
and UPPER(ch.HDBK) = 'VEN'
) results
ON aw.[Week] >= results.[Week]
WHERE aw.[Week] <= #MaxWeek
GROUP BY
aw.[Week]