Using functions or methods in Java's String.replaceAll() regex

Question

I am trying to convert an RPN-equation into a string that matches tigcc rules. There numbers must have the number of chars in front of them and a tag for positive or negative. For "2" it would be "1 2 POSINT_TAG"

My complete input to rpn converter is based on regexes, so I wanted to use them again and have a String.replaceAll() function like:

string.replaceAll("(\d+)","$1".length+" $1 POSINT_TAG");

But there it just prints: "2 number INT_TAG". I found some classes like com.stevesoft.pat (link).

Is there another way implemented in normal Sun Java to use (custom) functions in replace rules of regexes?

Alan Moore · Accepted Answer · 2012-07-27 17:13:03Z

13

No, at least not the same way you would do it in C# or Ruby.

The closest thing is to write a loop like this:

static Pattern pattern = Pattern.compile("\\d+");
String convert(String input) {
    StringBuffer output = new StringBuffer();
    Matcher matcher = pattern.matcher(input);
    while (matcher.find()) {
        String rep =
            String.format("%d %s POSINT_TAG",
                          matcher.group().length(),
                          matcher.group());
        matcher.appendReplacement(output, rep);
    }
    matcher.appendTail(output);
    return output.toString();
}

edited Jul 27, 2012 at 17:13

Alan Moore

75.6k13 gold badges110 silver badges161 bronze badges

answered Jan 20, 2011 at 0:26

finnw

48.8k24 gold badges150 silver badges223 bronze badges

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

3 Comments

Alan Moore Over a year ago

FYI, though StringBuilder is generally preferred, appendReplacement() and appendTail() do not accept it; you have to use StringBuffer instead. (StringBuilder did not exist yet when the Matcher class was added, in JDK 1.4.)

Ian Roberts Over a year ago

+1 to the comment - I keep hoping they'll add an overloading of these methods that can take any Appendable but (as of the current Java 8 builds) there's still only the StringBuffer version.

Air Over a year ago

You may be amused to know that this answer is cited in the source code for the Stack Overflow mobile app for Android.

Fred Chin · Accepted Answer · 2012-12-18 12:10:32Z

I have a idea about custom string replacement in Java. Not as easy as replacement function in JavaScript but it just works fine. Here is the code:


    package org.eve.util;

    import java.lang.reflect.Method;
    import java.lang.reflect.Modifier;
    import java.util.regex.Matcher;
    import java.util.regex.Pattern;

    public class StringUtils{
        public static String replaceWithFn(CharSequence input,String regexp,Method fn,int group){
            Matcher m=Pattern.compile(regexp).matcher(input);
            StringBuffer sb=new StringBuffer();
            try {
                Object obj=Modifier.toString(fn.getModifiers()).indexOf("static")>-1?
                        fn.getClass():fn.getDeclaringClass().newInstance();
                if(fn.getReturnType()!=String.class){
                    System.out.println("replacement function must return type \"String\".");
                }
                while(m.find()){
                    m.appendReplacement(sb, (String)fn.invoke(obj, m.group(group)));
                }
                m.appendTail(sb);
            } catch (Exception e) {
                e.printStackTrace();
            }
            return sb.toString();
        }
    }

    package org.eve.test;

import org.eve.util.StringUtils;

public class ReplaceTest {
    public static void main(String[] args) {
        try {
            StringBuffer input=new StringBuffer("\\u039D\\u03B9\\u03BA\\u03CC\\u03BB\\u03B1\\u03BF\\u03C2 Nicholas \\u5C3C\\u53E4\\u62C9\\u65AF");
            System.out.println("input="+input);
            String result=StringUtils.replaceWithFn(
                input,
                "\\\\u([0-9a-zA-Z]{4})",
                ReplaceTest.class.getMethod("hex2char",String.class),
                1
            );
            System.out.println("output="+result);
        } catch (SecurityException e) {
            e.printStackTrace();
        } catch (NoSuchMethodException e) {
            e.printStackTrace();
        }
    }

    public String hex2char(String s){
        //TODO replaceholder
        return String.valueOf((char)Integer.parseInt(s,16));
    }
}

Just for fun.

This works perfectly! I used an interface instead of reflection: github.com/redstone/LegacyFactions/blob/…

Alan Moore · Accepted Answer · 2013-11-26 17:58:38Z

1

Unfortunately what you are attempting here isn't going to work. Java uses applicative order evaluation meaning that the arguments are evaluated before the function is ever called. In your case you are getting the length of the two-character string "$1", not the number of digits that will be captured in group #1.

So when you use this line:

string.replaceAll("(\\d+)","$1".length+" $1 POSINT_TAG");

What the function sees is:

string.replaceAll("(\\d+)","2 $1 POSINT_TAG");

As for a solution to your problem, the answer posted by finnw will work.

edited Nov 26, 2013 at 17:58

Alan Moore

75.6k13 gold badges110 silver badges161 bronze badges

answered Jan 20, 2011 at 17:49

Bob9630

9539 silver badges20 bronze badges

Collectives™ on Stack Overflow

Using functions or methods in Java's String.replaceAll() regex

3 Answers 3

3 Comments

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related