How do I handle floating point value inputs for values greater than Float.MAX_VALUE

Question

I have an api which takes a number as a String input and i need to get the Float value of the number. I currently use the Float.ParseFloat method to get the float value of my String number.

According the java documentation of Float.ParseFloat, it doesn't mention anything about the input being greater than the Float.MAX_VALUE.

One of the ways I was thinking of doing this was by checking the length of the input String is greater than the length of the Float.MAX_VALUE.

Pls suggest how I can go about handling this.

Javier · Answer 1 · 2012-07-12T08:16:08.660

1

You can use greater precision. Try double or BigDecimal. There are also arbitrary precision libraries which are open.
- Here you can find how much each IEEE 754 format can hold: http://en.wikipedia.org/wiki/IEEE_754-2008 . Float would be near 1.234567*10^38
If you can't parse it properly (e.g. if there are too many significant digits or the exponent is too big: 1.23456789012345e5000) you won't be able either to hold it in a single precision float.

edited Jul 12 '12 at 08:16

answered Jul 12 '12 at 08:08

Javier

678
5
17

`BigDecimal` it is (Javier was too fast xD) – ioreskovic Jul 12 '12 at 08:10

score 1 · Answer 2 · answered Jul 12 '12 at 08:12

Although the javadoc doesn't make it clear, when I tested it, parseFloat of a String too large simply produced a Float of 'infinity'. You could use the isInfinite() method after creation to check the value.

Using something like BigDecimal would probably be a safer option here, especially if you'll be performing any arithmetic on your value.

score 0 · Answer 3 · edited Jul 12 '12 at 09:23

0

The exponent clips to the maximum exponent value. See the source, line 1197.

Perhaps check for some maximum useful value for your application?

edited Jul 12 '12 at 09:23

Aleksandr Kravets

5,750
7
53
72

answered Jul 12 '12 at 08:12

tbroberg

635
3
11

Actually it doesn't, that's only the first part of the process. See the answers by the people who have actually tried it. – user207421 Jul 12 '12 at 08:30
Ugh. Caught me. No idea where it converts that big honkin' exponent to infinity, but I'm leaving the comment for the sake of the pointer to the source. – tbroberg Jul 12 '12 at 09:39

score 0 · Answer 4 · answered Jul 12 '12 at 08:13

0

If the number is too big the result is set to Float.POSITIVE_INFINITY, as the rules of IEEE FP arithmetic require, and as a 10-second test shows.

answered Jul 12 '12 at 08:13

user207421

305,947
44
307
483

How do I handle floating point value inputs for values greater than Float.MAX_VALUE

4 Answers4