pass date dynamically to a query to calculate day wise results - sql-server

Question

I created a new topic as i am trying to divide my initial question into multiple pieces.
The initial topic can be found in the following link.
I have created a SQL Fiddle link with table and sample data.
This is the Query that i have right now.
SELECT
Order1,
COUNT(UNit.UNIT) AS Units,
SUM(CASE
WHEN (DATEDIFF(dd, INSV_DATE, '2015-01-21')) >= 31 THEN 31
WHEN (DATEDIFF(dd, INSV_DATE, '2015-01-21')) < 0 THEN 0
ELSE (DATEDIFF(dd, INSV_DATE, '2015-01-21'))
END) AS Days31
FROM UNIT
WHERE Unit.INSV_DATE < '2015-01-21'
AND UNIT.MODEL IN ('Toyota')
AND (UNIT.Customer IN ('Jona'))
GROUP BY [Order1],
customer
how do i loop and pass date dynamically in the Datediff for a period of one month?
I want the 31 days output calculated for day wise.
The output should be like
Date | Order1 | Unit | Day31
----------------------------------
May20 | 90909 | 5 | 128
May19 | 90909 | 4 | 124
May17 | 90909 | 2 | 62
I actually want to do something like the following.
SELECT
Order1,
COUNT(UNit.UNIT) AS Units,
SUM(CASE
WHEN (DATEDIFF(dd, INSV_DATE, '2015-05-20')) >= 31 THEN 31
WHEN (DATEDIFF(dd, INSV_DATE, '2015-05-20')) < 0 THEN 0
ELSE (DATEDIFF(dd, INSV_DATE, '2015-05-20'))
END) AS Days31
FROM UNIT
WHERE Unit.INSV_DATE < '2015-05-20'
AND UNIT.MODEL IN ('Toyota')
AND (UNIT.Customer IN ('Jona'))
GROUP BY [Order1],
customer
SELECT
Order1,
COUNT(UNit.UNIT) AS Units,
SUM(CASE
WHEN (DATEDIFF(dd, INSV_DATE, '2015-05-19')) >= 31 THEN 31
WHEN (DATEDIFF(dd, INSV_DATE, '2015-05-19')) < 0 THEN 0
ELSE (DATEDIFF(dd, INSV_DATE, '2015-05-19'))
END) AS Days31
FROM UNIT
WHERE Unit.INSV_DATE < '2015-05-19'
AND UNIT.MODEL IN ('Toyota')
AND (UNIT.Customer IN ('Jona'))
GROUP BY [Order1],
customer
SELECT
Order1,
COUNT(UNit.UNIT) AS Units,
SUM(CASE
WHEN (DATEDIFF(dd, INSV_DATE, '2015-05-18')) >= 31 THEN 31
WHEN (DATEDIFF(dd, INSV_DATE, '2015-05-18')) < 0 THEN 0
ELSE (DATEDIFF(dd, INSV_DATE, '2015-05-18'))
END) AS Days31
FROM UNIT
WHERE Unit.INSV_DATE < '2015-05-18'
AND UNIT.MODEL IN ('Toyota')
AND (UNIT.Customer IN ('Jona'))
GROUP BY [Order1],
customer
Running the same query for everyday with the different date.
If you can guide me in making a day wise query that will be great.

Answer 1

You can build the list of dates using a recursive CTE, then CROSS JOIN it against your table to get the list of days you want:
DECLARE #StartDate date = 'Jan 1, 2015'
DECLARE #EndDate date = DATEADD(DAY, 30, #StartDate)
;WITH cte AS (
SELECT #StartDate AS ReportDate
UNION ALL
SELECT DATEADD(DAY, 1, ReportDate)
FROM cte
WHERE ReportDate < #EndDate
)
SELECT Order1,COUNT(UNit.UNIT) As Units,sum(CASE
WHEN (datediff(dd,INSV_DATE,cte.ReportDate)) >= 31 THEN 31
WHEN (datediff(dd,INSV_DATE,cte.ReportDate)) < 0 THEN 0
ELSE (datediff(dd,INSV_DATE,cte.ReportDate))END) as Days31
FROM UNIT
CROSS JOIN cte
WHERE Unit.INSV_DATE < cte.ReportDate AND
UNIT.MODEL in('Toyota') AND(UNIT.Customer in('Jona' ))
group by [Order1],customer

Answer 2

Based on my understanding of what you want, here is my "guess":
DECLARE #startDate DATE = CAST(MONTH(GETDATE()) AS VARCHAR) + '/' + '01/' + + CAST(YEAR(GETDATE()) AS VARCHAR) -- cast as mm/dd/yyyy
DECLARE #endDate DATE = GETDATE() -- mm/dd/yyyy
--creates a list of dates for the running month
;WITH month_dates
AS (
SELECT [Date] = DATEADD(Day, Number, #startDate)
FROM master.dbo.spt_values
WHERE Type = 'P'
AND DATEADD(day, Number, #startDate) <= #endDate
)
SELECT month_dates.[Date]
,Order1
,COUNT(UNit.UNIT) AS Units
,sum(CASE
WHEN (datediff(dd, INSV_DATE, month_dates.[Date])) >= 31
THEN 31
WHEN (datediff(dd, INSV_DATE, month_dates.[Date])) < 0
THEN 0
ELSE (datediff(dd, INSV_DATE, month_dates.[Date]))
END) AS Days31
FROM UNIT
CROSS JOIN month_dates --cross join list of dates
WHERE Unit.INSV_DATE < month_dates.[Date]
AND UNIT.MODEL IN ('Toyota')
AND (UNIT.Customer IN ('Jona'))
AND UNIT.Order1 = 'A1056729' --added this filter to test the output
GROUP BY month_dates.[Date]
,[Order1]
,customer
ORDER BY month_dates.[Date] desc