Adding Double Variables Precision
Microsoft Visual C++ .NET 55510-640-8018297-18683
When I attempt to add two double type variables, I receive an incorrect result. For example:
double a,b,c,d;
//Test #1
a = 295.00;
b = 0.14;
c = a + b;
//Test #2
d = 14.00;
d /= 100;
When I debug, c = 295.139999999 and d = 0.13999999
These values should be 295.14 and 0.14, respectively. As a result, these unexpected results are causing my program to run incorrectly.
I have also tried to use rounding functions to set c and d correctly after they have been calculated, but have had no success.
Help on this issue would be greatly appreciated - Thanks.
[699 byte] By [
cgibson] at [2007-11-11 8:44:22]
# 2 Re: Adding Double Variables Precision
I guess I should add that this is occurring on the following software version:
Microsoft Development Environment 2002
Version 7.0.9955
Micorsoft .Net Framework 1.0
1.0.3705
# 3 Re: Adding Double Variables Precision
It is a rounding error due to the compiler's floating point implementation. I expect you can find information in MSDN for a work around.
http://msdn1.microsoft.com/en-us/default.aspx
nspils at 2007-11-11 21:03:05 >
# 4 Re: Adding Double Variables Precision
Print 'b' from the first example. I suspect you will see .13999999999999 because this value may be impossible to represent exactly in floating point notation. This should not break your program. Do you compare with an == operation? You should almost never use == on floating point numbers, instead something like
if ( fabs(x-y) <= .00000000001) or similar code.
jonnin at 2007-11-11 21:04:11 >
