in matcher.replace method,how to limit replace times?

Question

in matcher.replace method,only has: replaceFirst() and replaceAll() two methods

i want limit replace 3 times,how to do?

<

Answer 1

On appendReplacement/Tail

You'd have to use appendReplacement and appendTail explicitly. Unfortunately you have to use StringBuffer to do this. Here's a snippet (see also in ideone.com):

    String content="aaaaaaaaaa";
    Pattern pattern = Pattern.compile("a");
    Matcher m = pattern.matcher(content);

    StringBuffer sb = new StringBuffer();
    final int N = 3;
    for (int i = 0; i < N; i++) {
      if (m.find()) {
         m.appendReplacement(sb, "b");
      } else {
         break;
      }
    }
    m.appendTail(sb);
    System.out.println(sb); // bbbaaaaaaa

See also

• StringBuilder and StringBuffer in Java
  • StringBuffer is synchronized and therefore slower than StringBuilder
• BugID 5066679: Matcher should make more use of Appendable
  • If granted, this request for enhancement would allow Matcher to append to any Appendable

---

Another example: N times uppercase replacement

Here's another example that shows how appendReplacement/Tail can give you more control over replacement than replaceFirst/replaceAll:

// replaces up to N times with uppercase of matched text
static String replaceUppercase(int N, Matcher m) {
    StringBuffer sb = new StringBuffer();
    for (int i = 0; i < N; i++) {
        if (m.find()) {
            m.appendReplacement(
                sb,
                Matcher.quoteReplacement(m.group().toUpperCase())
            );
        } else {
            break;
        }
    }
    m.appendTail(sb);
    return sb.toString();
}

Then we can have (see also on ideone.com):

    Pattern p = Pattern.compile("<[^>]*>");
    Matcher m = p.matcher(" b c  e  g <$$$> i ");

    System.out.println(replaceUppercase(4, m));
    //  b c  e  g <$$$> i 
    //  1        2       3      4

The pattern <[^>]*> is just a simple example pattern that matches "".

Note that Matcher.quoteReplacement is necessary in this particular case, or else appending "<$$$>" as replacement would trigger IllegalArgumentException about an illegal group reference (because $ unescaped in replacement string is a backreference sigil).

---

On replaceFirst and replaceAll

Attached is the java.util.regex.Matcher code for replaceFirst and replaceAll (version 1.64 06/04/07). Note that it's done using essentially the same appendReplacement/Tail logic:

// Excerpt from  @(#)Matcher.java   1.64 06/04/07

public String replaceFirst(String replacement) {
    if (replacement == null)
        throw new NullPointerException("replacement");
    StringBuffer sb = new StringBuffer();
    reset(); // !!!!
    if (find())
        appendReplacement(sb, replacement);
    appendTail(sb);
    return sb.toString();
}

public String replaceAll(String replacement) {
    reset(); // !!!!
    boolean result = find();
    if (result) {
        StringBuffer sb = new StringBuffer();
        do {
            appendReplacement(sb, replacement);
            result = find();
        } while (result);
        appendTail(sb);
        return sb.toString();
    }
    return text.toString();
}

Note that the Matcher is reset() prior to any replaceFirst/All. Thus, simply calling replaceFirst 3 times would always get you the same result (see also on ideone.com):

        String content="aaaaaaaaaa";
        Pattern pattern = Pattern.compile("a");

        Matcher m = pattern.matcher(content);

        String result;

        result = m.replaceFirst("b"); // once!
        result = m.replaceFirst("b"); // twice!
        result = m.replaceFirst("b"); // one more for "good" measure!

        System.out.println(result);
        // baaaaaaaaa
        // i.e. THIS DOES NOT WORK!!!

See also

• java.util.regex.Matcher source code, OpenJDK version