Question

So just to give you an example of how our procedure would eventually work with data and come up with this error, here’s some SQL:

DECLARE @daysInPeriod  INT
DECLARE @dcFraction    DECIMAL(38,38)
DECLARE @YEAR_360         DECIMAL
SET @YEAR_360 = 360.0000000000
SET @daysInPeriod = DATEDIFF(dd, '12/24/1970', '6/29/07')
SET @dcFraction = @daysInPeriod/@YEAR_360

If you run this in SQL Server 2008 R2, you’ll see that you receive an 'Arithmetic overflow error converting numeric to data type numeric.' error.

The decimal that I am setting the value to is max’d out in terms of places it can hold, which I believe is 38. I don’t know what else I can do to get this value. Shouldn’t you be able to divide an integer with a decimal in SQL and get something out of it, especially if the decimal is max’d out?

Answer 1

In your example, the value of @daysInPeriod/@YEAR_360 is 37.0444444444444444444. So you can’t assign that value to a parameter of data type DECIMAL(38,38), since you are leaving no room the int part. A DECIMAL(38,38) means that you can store 38 digits, and 38 of them are on the decimal part, so any value greater than 0.999999999999 will throw an error.