regular expression

[Brian Buckley]

I want a regular expression to split a String on whitespace except not when the whitespace is within brackets.

So for example "The cat [in the] hat".split(regularExpression) should evaluate to "The","cat","[in the]","hat".

If the regularExpression were "\\s+" it works for whitespace alone. How can I change this to not split when the whitespace is in brackets?

Tips welcome. Thanks!

Brian

[Dirk Schreckmann]

Moving this to the Intermediate forum...

[Peter den Haan]

Use a zero-width (negative) look-behind (?<!) to verify that the space isn't preceded by a "[" without balancing "]":

(?<!\[[^\]]*)\s+

Be warned that I didn't try this, but it should be close. Regular expressions, the most useful write-only medium in existence Don't forget to escape the backslashes if you put this in a String literal.

- Peter

[Jim Yingst]

Mmmmm, I don't think that will work well, Peter. (Welcome back, BTW!) I get:

Exception in thread "main" java.util.regex.PatternSyntaxException: 
Look-behind group does not have an obvious maximum length near index 11

Java.util.regex only does lookbehind if it can determine a maximum length to what it's looking for. Other regex packages may only do it if the expression has a single fixed length. Friedl discusses this, but I don't have my copy with me right now.

It may be possible to adapt this pattern for the problem at hand, but I don't see an easy way to do it. I think it will be easier to write a pattern to match the non-whitespace (or anything enclosed in braces) instead. E.g.:

String noWhitespaceOrOpenBrace = "[^\\s\\[]++";
String matchingBraces = "\\[[^\\]]*+\\]";
String unmatchedBrace = "(\\[)"; // group 1
Pattern p = Pattern.compile(
        "(?:" + noWhitespaceOrOpenBrace + "|" + matchingBraces + "|" + unmatchedBrace + ")++");
String input = "fee fie    foe[] fum, [I smell the blood] [of an Englishman!]";
Matcher m = p.matcher(inputStr);
while (m.find()) { 
    if (m.group(1) == null)
        System.out.println("found token: " + m.group());
    else
        System.out.println("found an unmatched brace");
}
System.out.println("done");

[ June 05, 2004: Message edited by: Jim Yingst ]

[Max Habibi]

For the problem, as stated, the following should work.

 String splitRegex = " (?![^\\[\\]]*\\])";

However, it doesn't work with nested structures.

What we're saying here is the following. A space, that is followed any sequence of characters, so long as those characters are anything other then a open or closed bracket, but which end with a closed bracket


HTH,
M

[Brian Buckley]

My god that works. Regular expressions freak me out.

I have to sit down study this...

Brian

[Max Habibi]

I hear some idiot recently wrote a book on regex & Java.

M

[Peter den Haan]

And a good book on regex & Java to boot. What was his name again? Something that sounded a bit like a cool holiday destination. M... Max... Max... Oh, I don't know.

- Peter