NumberFormat and truncating fraction digits

[Justyna Wozniak]

Hello,
I am trying to find out why the following piece of code:

  
float fl = 123456.5668f;
NumberFormat nf = NumberFormat.getInstance();
System.out.println(nf.getMaximumFractionDigits());
System.out.println(nf.format(fl));
nf.setMaximumFractionDigits(4);
System.out.println(nf.format(fl));

displays this output:

  
3
123,456.57
123,456.5703
123456

If the default number of fraction digits is 3, why is .5668 truncated to .57?
And why does the code display .5703 when the number of fraction digits is extended to 4?

Thanks,
Justyna W.

[Dan Polak]

Hi it works very strange with floats

check
float fl = 123456.566f;

try to use double

bye

[Dan Polak]

123456.5668f is to big for float

[Chris Stann]

float gives 6-9 digits precision, although in your example I could only use 8 digits before truncation. Just add System.out.println(">"+fl+"<"); just under your float declaration to see.

Dealing with floats is tricky in general, as I have recently found out when attempting to do accounting arithmetic (think BigDecimal).

Use a double to get the expected result.

        double dl = 123456.5668d;
        NumberFormat nf = NumberFormat.getInstance();
        System.out.println(nf.getMaximumFractionDigits());
        System.out.println(nf.format(dl));
        nf.setMaximumFractionDigits(4);
        System.out.println(nf.format(dl));
    

result:

3
123,456.567
123,456.5668

Prosze bardzo!