SQL Server decimal variable assignment?

Question

I'm trying to write a stored procedure and I'm getting unexpected divide by 0 exceptions.

I've narrowed it down to the following example.

Why in the world does this :

    declare @A decimal;
    declare @B decimal;
    declare @C decimal;

    set @A = 4;
    set @B = 9;
    set @C = @A/@B 

    select @A/@B as 'Expected'
    select @C as 'Wut'

result in this?

    Expected
    ---------------------------------------
    0.4444444444444444444

    (1 row(s) affected)

    Wut
    ---------------------------------------
    0

    (1 row(s) affected)

Answer 1

The problem is that you haven't specified a scale for the decimal type. From MSDN:

s (scale)

The number of decimal digits that will be stored to the right of the decimal point. This number is substracted from p to determine the maximum number of digits to the left of the decimal point. Scale must be a value from 0 through p. Scale can be specified only if precision is specified. The default scale is 0; therefore, 0 <= s <= p.

So when you try to store @A/@B back into @C, the fractional part gets truncated.

Notice:

declare @A decimal(18, 3);
declare @B decimal(18, 3);
declare @C decimal(18, 3);

set @A = 4;
set @B = 9;
set @C = @A/@B 

select @A/@B -- 0.44444444444444444444
select @C    -- 0.444