issues when convert a double to a long

[SpeedBWild]

I am reading in a line of data so that I can insert the text data into an array. Here is my code:
double rate;
rate = atof (&wktxt[k]); /* &wktxt[k]=2.046 */
rate *= 1000.0;
prate[cin][x] = (long) rate;

The result of prate[cin][x] is 2045

I have hundreds of values in the text file all of them convert as expected except 2.046. If I add this line of
code rate += .00000000001; (see below) then it converts correctly.

rate *= 1000.0;
rate += .00000000001;
prate[cin][x] = (long) rate;

Any ideas? Has anyone else experienced anything like this before?

 

[cpjust]

If you do some research about floating point arithmetic you'll find your answer. Basically, it has to do with the way floating point numbers are stored...
Wikipedia should have lots of info:

http://en.wikipedia.org/wiki/Floating_point

 

[Salem]

If your intent is to store an int, then don't read it in as a float just because it's an easy way of dealing with the decimal point.

Read it in as two ints separated by a '.' and perform appropriate arithmetic to combine those two ints into a result.


--
If you dance barefoot on the broken glass of undefined behaviour, you've got to expect the occasional cut.

 

[ArkM]

Use some kind of rounding, for example:

prate[cin][x] = (long) (rate+0.5);

(to nearest value for positive numbers).