float contrast

[Sweta Pillai]

class test  {
public static void main(String [] args) {
 float f=10;                //---------1
 System.out.printf("%f", f);//----------2
 //System.out.printf("%f", 10);//----------3
 System.out.printf("%f", 10.0);//----------4
 //float f1=10.0;             // ------------5

}
}

1->OK
2->OK o/p =10.000000
3->java.util.IllegalFormatConversionException: f != java.lang.Integer
4->OK o/p =10.000000
5->Not OK (Type mismatch: cannot convert from double to float-No issue)

The question is: while 1 compiles, 5 doesn't. Ok. But while 3 causes exception, 4 using 10.0 does fine. Why this contrast between normal float usage and in format()/printf() ? I mean 10.0 acceptable in one case, while the same is not acceptable in other.

[Keith Lynn]

Originally posted by Sweta Pillai:

class test  {
public static void main(String [] args) {
 float f=10;                //---------1
 System.out.printf("%f", f);//----------2
 //System.out.printf("%f", 10);//----------3
 System.out.printf("%f", 10.0);//----------4
 //float f1=10.0;             // ------------5

}
}

1->OK
2->OK o/p =10.000000
3->java.util.IllegalFormatConversionException: f != java.lang.Integer
4->OK o/p =10.000000
5->Not OK (Type mismatch: cannot convert from double to float-No issue)

The question is: while 1 compiles, 5 doesn't. Ok. But while 3 causes exception, 4 using 10.0 does fine. Why this contrast between normal float usage and in format()/printf() ? I mean 10.0 acceptable in one case, while the same is not acceptable in other.

1 compiles because in Java, an int, which is the default type of 10, can be automatically promoted to a float.

However, 5 does not compile, because a double, which is the default type of 10.0, cannot be assigned to a float without an explicit cast.

In the format string, %f means a floating point number. So in 3, System.out.printf is expecting a float or a double. Since the type of the parameters of System.format are Object, primitives will be boxed to their corresponding wrapper classes. There is no relationship between Integer and Float or Integer and Double.

In 4, 10.0 is a floating point number so System.out.printf with a %f in the format string will accept 10.0.

[Raghavan Muthu]

Thats correct as Keith Lynn said.

When you simply assign an integer value to a floating point variable as

    float f = 10;
  

the integer value automatically gets promoted to float. Thats why you don't get any error on line 1 in your program.

Whereas, when you assign a value with a precision (decimal point), it means by default its of type double. So it CANNOT be assigned to a floating point variable which is of less size, without an explicit cast. Thats why you get an error on line 5 in your program.

You need to assign the value as

     float f = 10.0F;
   

Otherwise, the compiler would throw an error saying that "possible loss of precision".

HtH.

[Sweta Pillai]

Thanks! It's pretty clear now.