return a result even if there is no data for that day/ week/ month - sql

Question

I have an MSSQL database and I want to get a value for each day/ week/ month in separate queries.
I got this working just fine except for intervals where there is no data, it wont return anything. And since im putting this in a graph, I want it to display a 0 or a NULL at least instead of jumping days or weeks etc.
I dont know if it will be different for each query but here is my daily query:
select CAST(Placements.CreatedOn AS DATE) AS
date,SUM(Placements.CommissionPerc * (Placements.PlacementFee / 100)) AS value
from [placements]
where [Placements].[CreatedOn] >= '2018-06-07' and [Placements].[CreatedOn] < '2018-06-12'
group by CAST(Placements.CreatedOn AS DATE)
order by CAST(Placements.CreatedOn AS DATE) ASC
This returns a result like:
So it returns 0 for when the data is actually 0 but when its missing, theres nothing like for days 9, 10 and 12
How can i fix this? thanks

Answer 1

Using a recursive CTE you can generate a list of dates.
Which can then be used to LEFT JOIN your table.
Example:
WITH DATES2018 AS
(
SELECT CAST('2018-01-01' AS DATE) AS [date]
UNION ALL
SELECT DATEADD(day, 1, [date])
FROM DATES2018
WHERE [date] < CAST('2018-12-31' AS DATE)
)
SELECT
d.[Date],
SUM(p.CommissionPerc * (p.PlacementFee / 100.0)) AS [value]
FROM DATES2018 AS d
LEFT JOIN [Placements] AS p ON CAST(p.CreatedOn AS DATE) = d.[Date]
WHERE d.[Date] BETWEEN '2018-06-07' AND '2018-06-11'
GROUP BY d.[Date]
ORDER BY d.[Date] ASC
OPTION (MAXRECURSION 366)
But you could also just add a new permanent table with all dates.
And use that table to left join your table.
Btw, if variables are used for the start and end date then that SQL can be optimized.
DECLARE #StartDate DATE = '2018-06-07';
DECLARE #EndDate DATE = '2018-06-11';
WITH DATES AS
(
SELECT #StartDate AS [date]
UNION ALL
SELECT DATEADD(day, 1, [date])
FROM DATES
WHERE [date] < #EndDate
)
SELECT
d.[Date],
SUM(p.CommissionPerc * (p.PlacementFee / 100.0)) AS [value]
FROM DATES AS d
LEFT JOIN [Placements] AS p
ON p.CreatedOn BETWEEN CAST(#StartDate AS DATETIME) AND CAST(DATEADD(day, 1, #EndDate) AS DATETIME) AND
CAST(p.CreatedOn AS DATE) = d.[Date]
GROUP BY d.[Date]
ORDER BY d.[Date] ASC
OPTION (MAXRECURSION 0)

Answer 2

A permanent calendar table would be best but here's and example that uses a CTE to create the dates needed for a LEFT JOIN. This uses a maximum of 1,000 days but can be extended as needed.
DECLARE
#StartDate date = '2018-06-07'
, #EndDate date = '2018-06-12';
WITH
t10 AS (SELECT n FROM (VALUES(0),(0),(0),(0),(0),(0),(0),(0),(0),(0)) t(n))
,t1k AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT 0)) - 1 AS num FROM t10 AS a CROSS JOIN t10 AS b CROSS JOIN t10 AS c)
,calendar AS (SELECT DATEADD(day, num, #StartDate) AS calendar_date
FROM t1k
WHERE num <= DATEDIFF(day, #StartDate, #EndDate)
)
SELECT
calendar.calendar_date AS date
, SUM( COALESCE(Placements.CommissionPerc * (Placements.PlacementFee / 100),0 ) ) AS value
FROM calendar
LEFT JOIN [placements] ON [Placements].[CreatedOn] = calendar.calendar_date
GROUP BY calendar.calendar_date
ORDER BY calendar.calendar_date ASC;