Tricky RegEx Problem (Java)

dob99

I have a piece of code (Java) which matches arbitrary tokens in strings using regular expressions. I'm using the Apache RegEx library.

What I'd like to to do is craft a regular expression that only matches a token if the preceding character matches a certain criterion. The tricky part is that I don't want the preceding character to part of the match.

Here's the Java code:

public int match(RE regEx, String s) {
   int start = 0;
   while (regEx.match(s, start) {
      handleToken(regEx.getParen(0));
      start = regEx.getParenEnd(0);
   }
}

An example would be searching for printf-style placeholders. The following expression finds them alright:

%(( |-|\+|0|#)*([0-9]+|\*)?(\.([0-9]+|\*))?)[hlL]?[cCdiouxXeEfgGnpsS]

However, this would also match the "%s" in "%%s", which is incorrect. So, I'm looking for a way to match in "foo %s bar", but not in "foo %%s bar".

Obviously, I could do this in the Java code, but in order to keep it as flexible (and preferably simple) as possible, I'd like to do it using a regular expression only. This is especially important as not all expressions are going to have this requirement.

Kev

Does the Apache Regex Library you are using support negative lookbehind assertions ?

In perl I would use this at the start of your pattern

(?<!%)

dob99

Kev wrote: »

Does the Apache Regex Library you are using support negative lookbehind assertions ?

In perl I would use this at the start of your pattern

(?<!%)

That works with the built-in JDK Pattern class. Unfortunately it doesn't with Apache's one. Looks like I'll have to look into switching.

Thanks for the help.

Civilian_Target

Have you considered combining the tokenizer and the reg-ex expressions.

Certainly, I find this to be a handy way of doing replacements on XML files that are not tag based, use reg-exps to get to the right part and tokenize on some relevant character, like "