How do I get rid of xml tags but not when the tag has copyright in it?

Question

I tried this but it does not seem to do anything:
temp_string=temp_string.replaceAll("\\<.*?(!copyright)\\>", "");An example of this is that if I have: <meta name ="copyright" > I do not want to get rid of the tag. In other words, I do not want to get rid of any tags that have the word "copyright" in them, but I want to get rid of all other tags.
Edited by: setaret2004 on Dec 16, 2008 4:11 PM

Answer 1

setaret2004 wrote:
Also if I want to include license as well do I just do: str = str.replaceAll("<(?![^<>]*(copyright|license))[^<>]*>", ""); Is that correct?Yes, you can do the same thing with all three regexes. Now, explanations:
"<(?![^<>]*copyright)[^<>]*>" This regex does one lookahead, scanning forward for the word "copyright" before beginning the actual match. It's a negative lookahead, so if the word is seen, the lookahead fails and so does the overall match. If the lookahead succeeds, the match position is returned to where the lookahead started and the real matching begins.
A naïve approach would use ".&#x2A;" or ".&#x2A;?" in the lookahead, but then it would be free to scan past the end of the tag and find the word in the next tag or halfway through the document; lookaheads are slippery that way. But I used "[<>]&#x2A;", so once it reaches the closing ">" it stops scanning.
That means, for every successful match the regex scans the whole tag twice--once for the lookahead and once "for reals". That won't be a problem in most cases, but if you're matching many large chunks of text, performance could suffer. Each of the other regexes only traverses the tag once. "<(?:(?!copyright)[^<>])*>" This one steps through the text one character at a time, at each position doing a negative lookahead before matching the character. That sounds like a lot of work, but in the vast majority of cases the lookahead will only have to look at one character before yielding.
"<(?:(?!copyright|[<>]).)*>" This is basically the same as the second regex, but it does all the real work in the lookahead. I don't recommend it; it's less readable and probably less efficient than the second regex. But both the second and third regexes can be made more efficient by using [possessive quantifiers|http://www.regular-expressions.info/possessive.html] (that whole site is highly recommended, by the way): "<(?:(?!copyright)[^<>])*+>", "");
"<(?:(?!copyright|[<>]).)*+>" The first regex can use a possessive quantifier, too, but not in the lookahead: "<(?![^<>]*copyright)[^<>]*+>"