Java regexp overmatches

Question

I want to replace a single <, but not <<, using a Java regexp (so I'm working with String.replaceAll()). Right now I have

([^<]|^)<([^<]|$)

which works in Python, but not Java. I have also tried negative look-around, with something like

(?<!<)<(?!<)

However, all my current attempts match << as well. Am I doing something wrong? Is there any Java-specifics here I don't know about?

Answer 1

Perhaps you are using the wrong replacement?

\n is only used in backrefering to the capture group on the pattern side. You have to use $n to use it on the replacement side.

See http://java.sun.com/j2se/1.4.2/docs/api/java/util/regex/Matcher.html#replaceAll%28java.lang.String%29 for the documention on this.

Answer 2

Your first regex contains a character class ([^<]|^). That's a positive match, so whatever gets caught in the character class will be replaced by replaceAll().

Your second regex only uses lookaround, which only verifies a condition, and does not match:

(?<!<)<(?!<)

That one works fine for me: it does not match <<. Perhaps you can post a code snippet and some input/output that does not behave as you'd expect?

Answer 3

If you want to replace a single "<" with an "X" say, do this:

String test = "hello << world < blah < blah << blah";

String _test = test.replaceAll("(^|[^<])<([^<]|$)", "$1X$2");

System.out.println(_test);

Gives you this:

hello << world X blah X blah << blah

Edit Updated to match at beginning and end of line