Dynamic variable T-SQL (using stored procedure) - sql

Question

So I have got a table-valued function with parameters :
SampleProcedure(#date,#par1,#par2,#par3)
Date variable is an INT , for example :
#date int = 20170102
What I would like to do is to iterate through next days until EOF or specific , predefined date , so the #date variable should change once the previous iteration is done. Other parameters are not changing.
What approach should I take? I was wondering if I should use cursors , but I don't really understand them at the moment - I'd be thankful if anyone explains me them at this example (iteration through dates as ints).
EDIT :
More specific case :
I have got GetDailyUsageReal and GetDailyUsageForecast stored procedures.
GetDailyUsageReal(#date,#par1,#par2)
GetDailyUsageForecast(#date,#par1,#par2)
My input :
DECLARE #date int = 20170102,
#par1 INT = 4000,
#par2 INT = 1,
;WITH CTE as (SELECT Hour, SUM(CAST(UsReal AS DECIMAL(19, 6))) / 1000000 as Real, Day
FROM GetDailyUsageReal(#date,#par1,#par2)
Group BY Hour,Day),
CTE2 as (SELECT Hour, SUM(CAST(UsForecast AS DECIMAL(19, 6))) / 1000000 as Forecast, Day
FROM GetDailyUsageForecast(#date,#par1,#par2)
Group BY Hour,Day)
SELECT cte.Hour, Real, cte2.Forecast , cte.Day
FROM CTE
JOIN CTE2 on cte.hour=cte2.hour AND cte.day=cte2.day
ORDER BY cte.hour
The output is :
+------+------+----------+----------+--+
| Hour | Real | Forecast | Day | |
+------+------+----------+----------+--+
| 1 | 10 | 12 | 20170102 | |
| 5 | 24 | 23 | 20170102 | |
| 7 | 24 | 22 | 20170102 | |
| 8 | 27 | 27 | 20170102 | |
| 9 | 26 | 21 | 20170102 | |
| 10 | 21 | 21 | 20170102 | |
| 11 | 11 | 12 | 20170102 | |
| 12 | 25 | 24 | 20170102 | |
| 13 | 17 | 18 | 20170102 | |
| 14 | 18 | 19 | 20170102 | |
| 15 | 26 | 25 | 20170102 | |
| 16 | 22 | 21 | 20170102 | |
| 17 | 23 | 23 | 20170102 | |
| 18 | 24 | 23 | 20170102 | |
| 19 | 19 | 18 | 20170102 | |
| 20 | 10 | 11 | 20170102 | |
| 21 | 11 | 13 | 20170102 | |
| 22 | 18 | 16 | 20170102 | |
| 23 | 19 | 17 | 20170102 | |
| 24 | 11 | 13 | 20170102 | |
+------+------+----------+----------+--+
What I want to get is basically output for the next days, let's say until 2019 (there's some data even for 2019 in my DB).
So what I need is the iteration of date. I have no access to change #date data type to DATE.
#EDIT2 :
My expected output :
+------+------+----------+----------+--+
| Hour | Real | Forecast | Day | |
+------+------+----------+----------+--+
| 1 | 10 | 12 | 20170102 | |
| 5 | 24 | 23 | 20170102 | |
| 7 | 24 | 22 | 20170102 | |
| 8 | 27 | 27 | 20170102 | |
| 9 | 26 | 21 | 20170102 | |
| 10 | 21 | 21 | 20170102 | |
| 11 | 11 | 12 | 20170102 | |
| 12 | 25 | 24 | 20170102 | |
| 13 | 17 | 18 | 20170102 | |
| 14 | 18 | 19 | 20170102 | |
| 15 | 26 | 25 | 20170102 | |
| 16 | 22 | 21 | 20170102 | |
| 17 | 23 | 23 | 20170102 | |
| 18 | 24 | 23 | 20170102 | |
| 19 | 19 | 18 | 20170102 | |
| 20 | 10 | 11 | 20170102 | |
| 21 | 11 | 13 | 20170102 | |
| 22 | 18 | 16 | 20170102 | |
| 23 | 19 | 17 | 20170102 | |
| 24 | 11 | 13 | 20170102 | |
| 1 | 15 | 14 | 20170103 | |
| 5 | 18 | 11 | 20170103 | |
| 7 | 26 | 44 | 20170103 | |
| 8 | 21 | 33 | 20170103 | |
| 9 | 22 | 12 | 20170103 | |
| 10 | 21 | 21 | 20170103 | |
| 11 | 11 | 12 | 20170103 | |
| 12 | 15 | 12 | 20170103 | |
| 13 | 17 | 18 | 20170103 | |
| 14 | 18 | 19 | 20170103 | |
| 15 | 26 | 25 | 20170103 | |
| 16 | 22 | 21 | 20170103 | |
| 17 | 23 | 23 | 20170103 | |
| 18 | 24 | 23 | 20170103 | |
| 19 | 19 | 18 | 20170103 | |
| 20 | 10 | 11 | 20170103 | |
| 21 | 11 | 13 | 20170103 | |
| 22 | 18 | 16 | 20170103 | |
| 23 | 19 | 17 | 20170103 | |
| 24 | 11 | 13 | 20170103 | |
+------+------+----------+----------+--+
I just want to have values from dates between selected range ,or range from selected day till end of file - last row in DB basing on day (so the last day could be for example 20210131). I want to have them in one result table, as shown above.
#EDIT after changes :
Output :
+------+-----------+-----------+----------+
| Hour | Real | Forecast | Workdate |
+------+-----------+-----------+----------+
| 20 | 11.831587 | 15.140129 | 20170101 |
| 21 | 11.659364 | 15.003950 | 20170101 |
| 22 | 11.111199 | 14.736179 | 20170101 |
| 23 | 11.075579 | 14.812968 | 20170101 |
| NULL | NULL | NULL | NULL |
| 1 | 9.930323 | 12.856905 | 20170102 |
| 2 | 9.826946 | 12.741908 | 20170102 |
+------+-----------+-----------+----------+

Answer 1

#Pejczi, I have done this logic for you. You need a CTE to build all the dates that you are interested in. Then join the table function with an outer apply - this ensures that a valid date is passed to the function and thus returns the hour and forecast/real column for each date.
Let me know how it goes:
DECLARE #StartDate DATE='20170101'
DECLARE #EndDate DATE='20180601'--current_timestamp
DECLARE #Dates TABLE(
Workdate DATE Primary Key
)
;WITH Dates AS(
SELECT Workdate=#StartDate
UNION ALL
SELECT CurrDate=DateAdd(DAY,1,Workdate) FROM Dates WHERE Workdate<#EndDate
)
SELECT *
FROM
Dates D
OUTER APPLY
(
SELECT Hour, SUM(CAST(UsForecast AS DECIMAL(19, 6))) / 1000000 as Real, Day as WorkDate
FROM GetDailyUsageReal(CONVERT(CHAR(8),D.Workdate,112),#par1,#par2)
GROUP BY
Hour,Day
)R
OUTER APPLY
(
SELECT Hour, SUM(CAST(UsForecast AS DECIMAL(19, 6))) / 1000000 as Forecast, Day as WorkDate
FROM GetDailyUsageForecast(CONVERT(CHAR(8),D.Workdate,112),#par1,#par2)
GROUP BY
Hour,Day
)F
ORDER BY
d.Workdate
option (maxrecursion 0);

Answer 2

#openshac suggestion is valid. You should store date as DATE datatype and using StartDate/EndDate it will make it easier to query. See if you can replace:
DECLARE #DATE DATE ='20170102'
with
DECLARE #StartDate DATE ='20170102'
DECLARE #EndDate DATE ='20180102'

Answer 3

This version expands on the CTE dates and adds hour column, so you can join the real/forcasted table functions on the Hour.
DECLARE #StartDate DATETIME='20170101'
DECLARE #EndDate DATETIME='20170201'--current_timestamp
DECLARE #Dates TABLE(
Workdate DATE Primary Key
)
;WITH Dates AS(
SELECT Workdate=#StartDate,WorkHour=DATEPART(HOUR,#StartDate)+1
UNION ALL
SELECT CurrDate=DateAdd(HH,1,Workdate),DATEPART(HOUR,DateAdd(HH,1,Workdate))+1 FROM Dates WHERE Workdate<#EndDate
)
SELECT Workdate=CAST(Workdate AS date),WorkHour
FROM
Dates D
OUTER APPLY
(
SELECT Hour, SUM(CAST(UsForecast AS DECIMAL(19, 6))) / 1000000 as Real, Day as WorkDate
FROM GetDailyUsageReal(CONVERT(CHAR(8),D.Workdate,112),#par1,#par2) R
WHERE R.Hour=D.WorkHour
GROUP BY
Hour,Day
)R
OUTER APPLY
(
SELECT Hour, SUM(CAST(UsForecast AS DECIMAL(19, 6))) / 1000000 as Forecast, Day as WorkDate
FROM GetDailyUsageForecast(CONVERT(CHAR(8),D.Workdate,112),#par1,#par2) F
WHERE F.Hour=D.WorkHour
GROUP BY
Hour,Day
)F
option (maxrecursion 0);