为什么 f 放在 float 值之后？

By default 12.3 is double literal. To tell compiler to treat it as float explicitly -> use f or F.

See tutorial page on the primitive types.

float and double can only provide approximate representation values for some values. e.g. 12.3 or 0.1

The difference is that float is not as accurate (as it has less precision, because its smaller)

e.g.

System.out.println("0.1f == 0.1 is " + (0.1f == 0.1));
System.out.println("0.1f is actually " + new BigDecimal(0.1f));
System.out.println("0.1 is actually " + new BigDecimal(0.1));

prints

0.1f == 0.1 is false
0.1f is actually 0.100000001490116119384765625
0.1 is actually 0.1000000000000000055511151231257827021181583404541015625

So 0.1 is the closest representation in double and 0.1f is the closest representation in float

float fVariable = 12.3; is fine. but when you use only float value(without any identifier) in any expression that time you need to tell compiler that value is float hence we use suffix "f" after value. example

float fl =13f/5f;

here 13 and 5 are float values.

Seeing as there are only so many ways to represent a number in your program, the designers of Java had to cherry pick and assign each form to the most common use case. For those forms selected as default, the suffix that denotes the exact type is optional.

• For Integer literals (int, long), the default is int. For obvious reasons.
• For Floating point literals (float, double) the default is double. Because using double potentially allows safer arithmetic on the stored values.

So, when you type 12 in your program, thats an int literal, as opposed to 12L, which is a long. And when you type 12.3, thats a double literal, as opposed to 12.3F, which is a float.

So where is this relevant? Primarily in handling downcasts, or narrowing conversions. Whenever you downcast a long to an int, or a double to a float, the possibility for data loss exists. So, the compiler will force you to indicate that you really want to perform the narrowing conversion, by signaling a compile error for something like this:

float f = 12.3;

Because 12.3 represents a double, you have to explicitly cast it to a float (basically signing off on the narrowing conversion). Otherwise, you could indicate that the number is really a float, by using the correct suffix;

float f = 12.3f;

So too summarize, having to specify a suffix for longs and floats is a compromise the language designers chose, in order to balance the need to specify what exactly a number is, with the flexibility of converting numbers from one storage type to another.

Float is single-precision 32-bit IEEE 754 floating point and Double is double-precision 64-bit IEEE 754 floating point. When you use a value with decimal points and if you don`t specify is as 0.23f (specifically float) java identifies it as a double.

For decimal values, double data type is generally the default choice taken by java. Check This

During compilation, all floating point numbers (numbers with decimal point) default to double.

Therefore, if you don't want your number to double and just want it as float, you have to explicitly tell the compiler by adding a f or F at end of the literal constant.

In java we have many different basic variable types so in order to make it general , it has some default features. Like if we give an input 16.02 it automatically takes it as a double input. So if you want to specify it as float we mention that 'f' after the number or we can simply use:

float f = (float) 16.02;

or

float f = 16.02f;

In the same way we have to mention 16l if we want the number to be saved as a long type else it will automatically select the default type ie int type.

If you do not use f it will be interpreted as double, which is the default in Java. You can also write it like this##

float f=(float) 32.5956; float f=32.5956f;

[10 years after the original post]

why f or F is placed after float values

With the f, the initialization occurs with the closest float value. Without the f, it might differ.

Many code values like 12.3 are not exactly representable as a double or a float. Instead a nearby value is used.

One rounding

Code 12.3 converted to the closest float: 12.30000019073486328125

float fVariable1 = 12.3f;

Two roundings

Code 12.3 converted to the closest double 12.300000000000000710542735760100185871124267578125 and then to the nearest float: 12.30000019073486328125.

float fVariable2 = 12.3;

Sometimes that 2-step approach makes for a different value due to double rounding.