Java Cookbook

Rounding Floating-Point Numbers

Problem

You need to round floating-point numbers to integer or to a particular precision.

Solution

If you simply cast a floating value to an integer value, Java will truncate the value. A value like 3.999999 casted to an int or long will give 3, not 4. To round these properly, use Math.round( ) . There are two forms; if you give it a double, you get a long result. If you give it a float, you get an int.

What if you don’t like the rounding rules used by round? If you wanted to round numbers greater than 0.54 instead of the normal 0.5, you could write your own version of round( ):

// Round.java 
/** Round a number up if its fraction exceeds this threshold. */ 
public static final double THRESHOLD = 0.54; 
/* Return the closest long to the argument. 
 * ERROR CHECKING OMITTED. 
 */ 
static long round(double d) { 
    long di = (long)Math.floor(d);    // integral value below (or ==) d 
    if ((d - di) > THRESHOLD) 
        return di + 1; 
    else return di; 
}

If you need to display a number with less precision than it normally gets, you will probably want to use a DecimalFormat object.