processing float in java

Question

in java, i have an array of floats, some of them are negative numbers, like :

3.04, 9.02, -4.2, -3.21, 0.02, -0.34

I only want to get the float numbers higher than 2, so I used :

if(number > 2.0 ) print.. etc

but what i get is not only the numbers >2, also the negative numbers :

3.04, 9.02, -4.2, -3.21

what could be the problem ?

Could you post a more complete code snippet to show exactly what you are doing? — Péter Török
– Péter Török, Commented Jan 12, 2011 at 10:17
what could be the problem ? Looks to me like you took the absolute value of the numbers before the test. — Jeff Mercado
– Jeff Mercado, Commented Jan 12, 2011 at 10:20
the code is too large, actually i get the numbers from a file, each float number is written on a separate line, i read the line, and convert from string to float. and then apply the 'if' condition, could the problem be related to signed and unsigned numbers ? — newQuesion
– newQuesion, Commented Jan 12, 2011 at 10:23
Converting String to float... are you remembering to take the sign in as you parse the string? You should show us the code that assigns number and the part that prints it. — Jeff Mercado
– Jeff Mercado, Commented Jan 12, 2011 at 10:25

Jon Skeet · Accepted Answer · 2011-01-12 10:23:55Z

3

It's hard to tell what the problem is without seeing code. Writing the obvious code from your description, it works fine:

public class Test {
    public static void main(String[] args) {
        float[] floats = { 3.04f, 9.02f, -4.2f, -3.21f, 0.02f, -0.34f };

        for (float number : floats) {
            if (number > 2.0) {
                System.out.println(number);
            }
        }
    }
}

Compare your code with mine - if you still can't get yours to work, please edit your question with your code.

answered Jan 12, 2011 at 10:23

Jon Skeet

1.5m893 gold badges9.3k silver badges9.3k bronze badges

Sign up to request clarification or add additional context in comments.

Comments

darioo · Accepted Answer · 2011-01-12 10:49:19Z

Based on comments and question, assuming we have a file named numbers.txt with content:

3.04 
9.02 
-4.2 
-3.21 
0.02 
-0.34

this code should do the trick (I ommited exception handling for readibility):

float[] floats = new float[100];  // adjust as needed

String fileName = "numbers.txt";
File f = new File(fileName);
BufferedReader br = new BufferedReader(new FileReader(f));

String line = "";

int i = 0;

while( (line = br.readLine()) != null) {

    if(line == null || line.isEmpty()) continue;  // just in case there are empty lines in file
    floats[i] = Float.parseFloat(line);
    i++;
}

br.close();

for (float number : floats) {
    if (number > 2.0) {
        System.out.println(number);
    }
}

Output:

3.04
9.02

Gugussee · Accepted Answer · 2011-01-12 11:00:20Z

-1

This is too long for a comment but it should be pointed out that if you write:

if (number > 2.0)

then you probably shouldn't be working with floating-point numbers because you don't understand yet how they work.

There's a good Wikipedia page on floating point numbers:

http://en.wikipedia.org/wiki/Floating_point

Then I suggest reading the 80 pages long "What every computer scientist should know about floating-point numbers".

Basically: due to the limited precision of floating point numbers, comparison should typically always be done using an epsilon.

Things like: if (number > 2.0) are a big no-no.

answered Jan 12, 2011 at 11:00

Gugussee

1,7233 gold badges17 silver badges26 bronze badges

3 Comments

Péter Török Over a year ago

I don't fully agree. Equality tests are indeed a no-no for floating point values, but < and > is usually not so critical (this of course depends on the domain, the range of values etc.). And overall, this is definitely not an answer to the original question - it would have been better to squeeze it into a comment.

Gugussee Over a year ago

@Péter Török: which is why I started with "This is too long for a comment". Equality are a no-no that is a given. And > and < test have to be done using an epsilon and this is critical (or you get "Bad Things" [TM] like space shuttle blowing in the sky ; ) Btw I could give trivial examples where a < or > test without an epsilon will fail but this can be as trivially Googl'ed :)

Péter Török Over a year ago

would be quite interested to know which space shuttle blew up due to a floating point comparison without epsilon ;-) I understand how such tests can fail - my point was that < and > comparisons are rarely so critical in real life. (And if they are, one should use a BigDecimal anyway.)

Collectives™ on Stack Overflow

processing float in java

3 Answers 3

Comments

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related