Getting rid of commas in large numbers?

[wwuster]

Can someone tell me a regular expression to get rid of commas in a large number using
String.replaceAll()?
For example, change "100,000" to "100000"

thanks,
William

[Fubarable]

I would try to make this Locale-specific by using a NumberFormat object. This way you could easily change this to be able to accommodate other ways of writing numbers (for instance Spain where they use a decimal point where you have a comma here). Something like so:

import java.text.NumberFormat;
import java.text.ParseException;
import java.util.Locale;

public class Fubar001
{
  public static void main(String[] args)
  {
    String s1 = "100,000";
    String s2 = "";
    NumberFormat nf = NumberFormat.getInstance(Locale.getDefault()); // make locale-specific
    try
    {
      int i1 = nf.parse(s1).intValue();
      s2 = String.valueOf(i1);
      System.out.println(s2);
    }
    catch (ParseException e)
    {
      e.printStackTrace();
    }
  }
}

[wwuster]

These numbers with commas are embedded in existing strings that I didn't create and could be anywhere in the strings, so I need a general way to remove the commas in numbers.

thanks,
William

[Fubarable]

so I need a general way to remove the commas in numbers.

So let's first see your code that demonstrates your attempt to solve this, and then we'll then know where to go with this.

[wwuster]

So let's first see your code that demonstrates your attempt to solve this, and then we'll then know where to go with this.

I'd like to simply use a regular expression in String.replaceAll.

String str = "there are 123,000 blue items and 117,287 red items";
String regex = "???????";
str = str.replaceAll(regex, str);

After this, str would be "there are 123000 blue items and 117287 red items"

I'm just learning regular expressions and don't know how to create such a regular expression. That's my question.

William

[Fubarable]

I'm just learning regex myself, but I think that to solve this you'll need both look aheads and look behinds. You may wish to try this regex String:

String regex = "(?<=[\\d])(,)(?=[\\d])";

something like:

import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class Fubar002
{
  public static void main(String[] args)
  {
    String regex = "(?<=[\\d])(,)(?=[\\d])";
    Pattern p = Pattern.compile(regex);
    
    String str = "there are 123,000 blue items and 117,287 red items";
    Matcher m = p.matcher(str);

    str = m.replaceAll("");
    System.out.println(str);
  }
}

To Darryl.Burke: Hey, I may be able to learn some of this regex stuff after all!

[wwuster]

Thanks, that works. I need to learn about look aheads now that I know what's involved.

William

[Norm]

I took a slightly different approach. I used capture groups to repeatedly scan the String for text before the , the , and text after the ,. The before text must end with digit and the after text must start with 3 digits.

	public static void main(String[] args) {
		final String data = "This short, quick test for removing , from 1,235,789 and more.";

      Pattern pattern = Pattern.compile("(.*?\\d+)(,)(\\d{3}.*?)"); 
        // Defines 3 groups (in parans):
        // 1 - before the ,
        // 2 - the ,
        // 3 - after the comma
      
      String newData = data;           // Prime the work variable

      while(true) {
         Matcher matcher = pattern.matcher(newData);
         if(matcher.find()) {
            int gc = matcher.groupCount();
            for(int i=0; i <= gc; i++) {
               System.out.println("group" + i + "=" + matcher.group(i)); 
               //group0=This short, quick test for removing , from 1,235
               //group1=This short, quick test for removing , from 1
               //group2=,
               //group3=235
            }
            newData = matcher.replaceFirst("$1$3"); 
            System.out.println(" new=" + newData + "\ndata=" + data);
            // new=This short, quick test for removing , from 1235 and more.
            //data=This short, quick test for removing , from 1,235 and more.
            // new=This short, quick test for removing , from 1235789 and more.  <<< Final result
            //data=This short, quick test for removing , from 1,235,789 and more. 

   
         }else{
            break;                     // Exit loop when no match found
         }
      } // end while()
	} // end main

[Darryl.Burke]

I like keeping it as simple as possible. The only condition to match here is a comma preceded and follwed by a digit character.

This is alomost the same as the solution already posted by Fubarable, but uses String#replaceAll which calls the same Pattern and Matcher methods under the hood.

public class RemoveComma {

   public static void main(String[] args) {
      String regex = "(?<=\\d),(?=\\d)";
      String input = "there are 123,000 blue items and 117,287 red items";

      String output = input.replaceAll(regex, "");
      System.out.println(output);
   }
}

db

[Fubarable]

Thanks Darryl, you're the man!