Primitive float to int conversion is possible but same is not possible for reference type, why?

Question

This may be the simple answer but I want to know the detail.

// Narrowing Conversion float ff = Integer.MAX_VALUE; int ii = (int)ff; System.out.println("float: " + ff + " int: " + ii); // Reference Conversion Integer integer; Float floatt = 100F; integer = (Integer) floatt;

In the above example the narrowing conversion works properly but the same with reference conversion is not working. Detail explanation is highly appreciable.

It is not possible because Integer does not inherit from Float. Autocasting on object does only work from sub- to superclasses, no exception. Explicit casting is not possible because Integer and Float do not have direct connection (the common superclass is not enough for typecasting). — Turing85
– Turing85, Commented Jul 19, 2015 at 3:17
Ok make sense. So for all the numeric type direct conversion is only possible to Number type. Number number = floatt // because Number is a super type for Float class. — Bhuwan Gautam
– Bhuwan Gautam, Commented Jul 19, 2015 at 3:29
Just to be practical, you can use floatt.intValue() to convert a float to an int. — markspace
– markspace, Commented Jul 19, 2015 at 3:36
I saw you asked a lot of questions without accepting one. If some answer help you, just click the green mark beside it — phuclv
– phuclv, Commented Jul 19, 2015 at 6:15

krock · Accepted Answer · 2015-07-19 03:16:59Z

You can cast floats to ints, but you cannot cast the wrapper objects to different types. You must first unwrap the Float wrapper and then you can cast the primitive float to an int.

Java will auto box/unbox primitives and their wrapper types, java will also widen primitive types automatically. But it won't do both.

phuclv · Accepted Answer · 2015-07-19 04:20:04Z

Let's see the bytecode for details. Use javap -c to see the byte codes. The code

float ff = Integer.INT_MAX; int ii = (int)ff;

compiles to

 0: ldc #3 // float 2.14748365E9f 2: fstore_1 3: fload_1 4: f2i 5: istore_2

If it's primitive type cast, as you can see from bytecode, f2i is inserted by the compiler. (as line 4 shows).

On the other hand, if you use wrapper class, compiler doesn't insert any type conversion instruction.

That's why you cannot cast them.

Collectives™ on Stack Overflow

Primitive float to int conversion is possible but same is not possible for reference type, why?

2 Answers 2

Comments

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Related