Floating Point

When Java sees a numeric literal with a ., such as 0.0, it assumes it is seeing a double, and range is less important.

• A number in scientific notation is considered a double: 6.022e23 (6.022 * 10²³).

• Because of this, the following declarations are illegal:

float f = 0.0; // WILL NOT COMPILE
float sci = 3e10; // NO, e10 is a double.

• The reason is that 0.0 is a double, and a double (8 bytes) will not fit into a float (4 bytes).

We can tell Java a literal is a float by adding F or f after the number (case does not matter.)

float f1 = 0.0f;
float sciF = 3e10F; // This works.

We can even assign int literals to float or double values. * This works because int and float are both 4 bytes, and putting an int into a float will not lose any precision (like dropping decimal places).

// f after an int literal
float f2 = 1f;
// no F, but it works
float f3 = 1;

Practice Exercise

We can assign many whole number values to floats, and add F after most of them - but not binary numbers.

float f4 = 0b10101;  // binary number - an int
float f5 = 0b10101f; // NO, don't use F with binary numbers
float f6 = 0xF16F;   // F doesn't do anything here, it is a hex digit
float f7 = 0761;     // octal number
float f8 = 0761F;    // octal number with F

Drill

DataTypes/com.example.datatypes.drills.FloatValues * Copy the declarations of f1 to f8 into main and see the errors for yourself.

---

Prev -- Up -- Next