# Problem with DecimalFormat

I have this odd problem with DecimalFormat

...
private DecimalFormat f = new DecimalFormat("00.00");

public void paint(Graphics g)
{
g.drawString(f.format(1.15),10,30);
g.drawString(f.format(1.20),10,40);
g.drawString(f.format(1.50),10,50);
}
...

Above draws "01,14", "01,20" and "01,50" on the applet ???
As you can see the first drawString is wrong (I am missing 0.01), if I just draw the numbers without the format it works fine?

I have tryed with other numbers but it is only when I try to draw the figures from 1.13 to 1.15 I have this problem.
###### Who is Participating?

Commented:
Try this code below:

DecimalFormat f = new DecimalFormat("00.00");
double a=1.15;
System.out.println(f.format(a));

you will get 1.14... It's the general problem with double/float calculation.

The problem is in the statement "double a=1.15"
In fact, a is 1.149999999999... in internal process!
So, the '0.0099999999...' will be truncated by DecimalFormat.
I don't know whether it is bug or not in Java.

Here is two way to fix this 'bug??':
1. if you want to get the 'flood' of the double value:
double p=0.000000000001d;
String s=f.format(a+p);
2. if you want to get the 'neighborhood' of the double value with the foramt '00.00' :
double p=0.000000000001d;
String s=f.format(a+0.005d-p);

hope it's useful.
0

Author Commented:
Edited text of question
0

Commented:
sorry, my English is not good...

There are 2 ways to fix this 'bug??'....
0

Author Commented:
This must be a bug in the DecimalFormat, as this problem only occurs when using DecimalFormat, and the problem only occures with figures in the range from 1,13 to 1,15.

Your first solution seems useable for my applet, but is it possible that I will get new problems in another range of figures when adding 0.000000000001d ? (will other figures than above will be rounded up) ?
0

Commented:
All the double/float numbers will be stored as binary data in byte code.
There is no binary data to present as 1.15 exactly. System just chooses the most near data(1.14999999...) to be 1.15.
Because there is no difference between 1.149999.... and 1.15, DecimalFormat can't know that the 1.1499999... is from 1.15 originally.
Of cause, It will happen in any possible numbers and any System(including java,c++,vb...).
I think it's not a bug in Java, It's a 'problem' in Programming for double/float calculation.

For your case, you can use the code simply:
String s=f.format(a+0.005d);  // a is any double
It's can get the neighborhood of 'a' for display.

Or You can add a suitable precision (p=0.00000001) for any double when you display it.

Maybe, you should use java.math.BigDecimal, if you don't care the slow speed, it's a number format for any precision.
0

Commented:
If you think that to theat 2.99999999 as 3.0 is OK, you can just add the precision factor (p=0.00000001) to the variable when you display it or transfer it into interger. That's why I call the 'p' as precision factor.
And, when you calculate with double, you don't need to add p into it.
Then, It effects the printing of variable only. There won't  be any other problems generated by it.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.