Count differnce of two Date's with Respect that every month has EXACT 30 days - tsql

Question

I thought it was clear, but doesn't seem so.
This question is about T-SQL (since it's tagged with tsql :) )
So I couldn't find any out-of-the-box solution to calculate my problem.
Let's assume you have these two dateTimes:
DECLARE #start DATETIME = '2011-01-01',
#end DATETIME = '2011-04-15'
The difference of these two datetimes in Days should be quivalent to 105.
The calculation works as follows: For every full month add 30 days, for the rest add the days till the date is achieved.
I could program this, but it would be an enormous SQL-statement, which I find find kinda ugly.
Is there any simple solution for this, like a built-in function or something short?
Thanks in advance.

Answer 1

Does this do the trick?
;with dates as
(
SELECT
CAST ('2011-01-01' AS DATETIME) as start_date
,CAST('2011-04-15' AS DATETIME) as end_date
)
SELECT
start_date
,end_date
,CASE WHEN DATEDIFF(MM,start_date,end_date) = 0 THEN DAY(end_date) - DAY(start_date)
WHEN DAY(start_date) = 1 THEN (30 * (DATEDIFF(MM,start_date,end_date))) + DAY(end_date)
WHEN DAY(start_date) <> 1 THEN 30 * DATEDIFF(MM,start_date,end_date) + (DAY(end_date) - DAY(start_date))
END AS gap_in_days
FROM dates

Answer 2

Short Answer
There's no built in function, but you could pretty easily create your own to handle converting a datetime to an int. From there, the SQL you would have to write would be trivial.
Long Answer
There's no built in function that will do this, probably because every month doesn't have 30 days. :)
You can start with this:
DECLARE #start DATETIME = '2011-01-01',
#end DATETIME = '2011-04-15'
DECLARE #endConverted INT
SELECT #endConverted = DATEPART(month, #end) * 30
+ CASE
WHEN DATEPART(DAY, #end) <= 30
THEN datepart(DAY, #end)
ELSE 30
END
DECLARE #startConverted INT
SELECT #startConverted = DATEPART(MONTH, #start) * 30
+ CASE
WHEN DATEPART(DAY, #start) <= 30
THEN DATEPART(DAY, #start)
ELSE 30
END
SELECT #endConverted - #startConverted
This isn't beautiful SQL, but it works. Note that it returns 104 (because 15 days - 1 day = 14 days), but simple enough to tack on a + 1 to the end of the final select if you want to handle the boundry days differently.
Note that the math here could pretty easily be moved into a function, which would allow you to clean your SQL up. Let's assume you created a function called GetDateTimeAsInt which holds the math; your SQL could be as simple as
DECLARE #start INT = GetDateTimeAsInt('2011-01-01'),
#end INT = GetDateTimeAsInt('2011-04-15')
SELECT #end - #start -- may need to add 1 here
In my testing, this seems to work. It will return the same result as the DATEDIFF function for the date range you specify in your post, but this is because there are 2 days with 31 days and 1 day with 28, so effectively, Jan - April have 30 days each. If you use it with a wider date range, you'll begin to get different results with my code vs. the DATEDIFF function.
Hope this helps.

Answer 3

I use PERIODDIFF. To get the year and the month of the date, I use the function EXTRACT:
SELECT PERIOD_DIFF(EXTRACT(YEAR_MONTH FROM NOW()), EXTRACT(YEAR_MONTH FROM time)) AS months FROM your_table;
T-sql
SELECT DATEDIFF(dd, "2011-01-01","2011-04-15")