Java Regex to replace string surrounded by non alphanumeric characters

Question

I need a way to replace words in sentences so for example, "hi, something". I need to replace it with "hello, something". str.replaceAll("hi", "hello") gives me "hello, somethellong".

I've also tried str.replaceAll(".*\\W.*" + "hi" + ".*\\W.*", "hello"), which I saw on another solution on here however that doesn't seem to work either.

What's the best way to achieve this so I only replace words not surrounded by other alphanumeric characters?

Mark Peters · Accepted Answer · 2012-04-28T16:45:07.560

4

Word boundaries should serve you well in this case (and IMO are the better solution). A more general method is to use negative lookahead and lookbehind:

 String input = "ab, abc, cab";
 String output = input.replaceAll("(?<!\\w)ab(?!\\w)", "xx");
 System.out.println(output); //xx, abc, cab

This searches for occurrences "ab" that are not preceded or followed by another word character. You can swap out "\w" for any regex (well, with practical limitations as regex engines don't allow unbounded lookaround).

edited Apr 28 '12 at 16:45

answered Apr 28 '12 at 16:30

Mark Peters

80,126
17
159
190

Very nice! 1+ I'm betting that you are familiar with the tutorial site I've linked to. – Hovercraft Full Of Eels Apr 28 '12 at 16:32
@Hovercraft: It's open in another tab right now :P. But that's because I always forget the syntax of lookahead/lookbehind. – Mark Peters Apr 28 '12 at 16:34

Hovercraft Full Of Eels · Answer 2 · 2012-04-28T16:35:57.473

2

Use \\b for word boundaries:

String regex = "\\bhi\\b";

e.g.,

  String text = "hi, something";
  String regex = "\\bhi\\b";
  String newString = text.replaceAll(regex, "hello");

  System.out.println(newString);

If you're going to be doing any amount of regular expressions, make this Regular Expressions Tutorial your new best friend. I can't recommend it too highly!

edited Apr 28 '12 at 16:35

answered Apr 28 '12 at 16:28

Hovercraft Full Of Eels

283,665
25
256
373

i cut and paste the code exactly and this is the output "hi, something" – Bruce Lowe Apr 28 '12 at 16:33
1

You want `replaceAll`, not `replace`, as the latter works on literals, not regexes. – Mark Peters Apr 28 '12 at 16:34
1

changing to replaceAll as Mark suggested fixes the issue – Bruce Lowe Apr 28 '12 at 16:36
@Bruce: dang fat fingers. `replace` changed for `replaceAll` – Hovercraft Full Of Eels Apr 28 '12 at 16:36
otherwise, nice one, i personally learnt something new today +1 :) – Bruce Lowe Apr 28 '12 at 16:37

Java Regex to replace string surrounded by non alphanumeric characters

2 Answers2

Linked