You have a number of a larger type and you want to store it in a variable of a smaller type.

Cast the number to the smaller type. (A cast is a type listed in parentheses before a value that causes the value to be treated as though it were of the listed type.)

For example, to cast a long to an int, you need a cast. To cast a double to a float, you also need a cast.

This causes newcomers some grief, as the default type for a number with a decimal point is double, not float. So code like:

float f = 3.0;

won’t even compile! It’s as if you had written:

double tmp = 3.0; 
float f = tmp;

You can fix it either by making f be a double, by making the 3.0 be a float, by putting in a cast, or by assigning an integer value of 3:

double f = 3.0; 
float f = 3.0f; 
float f = 3f; 
float f = (float)3.0; 
float f = 3;

The same applies when storing an int into a short, char, or byte:

// CastNeeded.java 
public static void main(String argv[]) { 
    int i, k; 
    double j = 2.75; 
    i = j;            // EXPECT COMPILE ERROR 
    i = (int)j;        // with cast; i gets 2 
    System.out.println("i =" + i); 
    byte b; 
    b = i;            // EXPECT COMPILE ERROR 
    b = (byte)i;    // with cast, i gets 2 
    System.out.println("b =" + b); 
}

The lines marked EXPECT COMPILE ERROR will not compile unless either commented out or changed to be correct. The lines marked “with cast” show the correct forms.