Weird round off error in C program

[amishera]

Hi,

I have this piece of code in a C program where I trying to convert a double to a long long:

double x =  0.507462687;

double a = pow(10,k+j);
double b = pow(10,k);

double c = a*x;
double d = (long long) (b*x);
long long num,denum;

num = (long long) (c-d); // the value of num becomes = ........418 instead of ....419 ie the last digit 9 becomes 8.

I am not sure what I should do to make the thing working ie avoid this peculiar round off error. Please suggest something.

Thanks.

[JosAH]

Read this: Floating Point Arithmetic.

kind regards,

Jos