RE: Decimal rounding of a float data type in Java.

["Clapham, Paul" <pclapham@xxxxxxxxxxxxx>]

I expect you didn't notice the original source of this thread, where the
questioner was asking why he or she got something like 8.23999999999993755
instead of 8.24, after a simple calculation.  This IS how floating point
numbers work, regardless of your innumerate denial.  Floating point numbers
work in binary, not in decimal.  So dividing by 100 is not "just a two-digit
shift" in floating point, that would be dividing by 4.  And by the way,
323/34 is no more irrational than 422/100.  Both are rational numbers.

-----Original Message-----
From: James Donkin [mailto:James@DuMaurier.co.uk]
Sent: March 2, 2001 01:25
To: JAVA400-L@midrange.com
Subject: Re: Decimal rounding of a float data type in Java.



> Well, not exactly.  Your "average" doesn't become 4.22, it becomes
something
> that's extremely close to 4.22.  Floating point numbers can't express
*any*
> fraction exactly unless it's a power of 0.5.  So don't be surprised if you
> do "System.out.println(average)" and you find something like
> "4.2200000000033775" appearing.
e.g. average = 422 / 100;

This will always give 4.22. The time you get problems is with 323 / 34 or
some other fraction which generates an irrational number. Saying that

    float average = 422 / 100;

 will give 4.22000032 is like saying that

    float average = 4.22;

 will leave average = 4.2200042. Dividing by 100 just generates a two digit
shift.

+---
| This is the JAVA/400 Mailing List!
| To submit a new message, send your mail to JAVA400-L@midrange.com.
| To subscribe to this list send email to JAVA400-L-SUB@midrange.com.
| To unsubscribe from this list send email to JAVA400-L-UNSUB@midrange.com.
| Questions should be directed to the list owner: joe@zappie.net
+---