Representing float values in Java

Question

Look at the three lines of code below.

  float f = 1;

  float g = 1.1;

  float h = 1.1f;

Second line has compilation errors, while the other lines do not have compilation errors. First line is working fine without suffix f and third line is working with suffix f. Why is this?

Solution 1

Floating point literals in Java is a double value by default.

JLS 3.10.2 Floating-Point Literals

A floating-point literal is of type float if it is suffixed with an ASCII letter F or f; otherwise its type is double and it can optionally be suffixed with an ASCII letter D or d.

You can't assign a double value to a float without an explicit narrowing conversion. You therefore have two options:

• For literals, use the suffix f or F to denote a float value
• For non-literals, use an explicit cast (float)

An example of the latter is:

double d = 1.1;
float f = (float) d; // compiles fine!

---

On widening conversions

The reason why this compiles:

float f = 1;

is because the widening conversion from int to float can be done implicitly in the context of an assignment.

JLS 5.2 Assignment Conversion

Assignment conversion occurs when the value of an expression is assigned to a variable: the type of the expression must be converted to the type of the variable. Assignment contexts allow the use of one of the following:

• a widening primitive conversion (§5.1.2)
• [...]

JLS 5.1.2 Widening Primitive Conversion

The following 19 specific conversions on primitive types are called the widening primitive conversions:

• int to long, float, or double
• [...]

---

Other data type suffix for literals

As mentioned above, there's also the D or d suffix for double. Consider this snippet for example:

static void f(int i) {
    System.out.println("(int)");
}
static void f(double d) {
    System.out.println("(double)");
}

//...
f(1);   // prints "(int)"
f(1D);  // prints "(double)"

There's also a suffix for long literals, which is L or l (lowercase letter). It is highly recommended that you use the uppercase variant.

JLS 3.10.1 Integer Literals

An integer literal is of type long if it is suffixed with an ASCII letter L or l (ell); otherwise it is of type int. The suffix L is preferred, because the letter l (ell) is often hard to distinguish from the digit 1 (one).

Solution 2

You're assigning a double value to a float variable. 1.1 by itself (without the f tacked on the end) is assumed by the compiler to be of type double. The compiler doesn't like to make implicit downcasts because there's potential there to lose precision.

Solution 3

First line automatically casts int to float (ok).

Second line could not cast double to float because of loss of precision. You can make an explicit cast:

float g = (float) 1.1;

Third line does not need modification.

Comments

[Amber]

(To be more explicit - floating-point constants in Java are double by default, unless they include the f on the end which makes them float.)