How to exclude occurrence of a substring from a string using regex?

Question

I have a string input in the following two forms.

1.

<!--XYZdfdjf., 15456, hdfv.4002-->
<!DOCTYPE

2.

<!--XYZdfdjf., 15456, hdfv.4002
<!DOCTYPE

I want to return a match if the form 2 is encountered and no match for the form 1. Thus basically I want a regex that accepts arbitrarily all characters between <!-- and <!DOCTYPE , except when there is an occurance of --> in between.

I am using Pattern, Matcher and java regex. Help is sought in terms of a regex specifically usable with Pattern.compile()

Thanks in advance.

Answer 1

Pattern p = Pattern.compile("(?s)<!--(?:(?!-->).)*<!DOCTYPE");

(?:(?.-->).)* matches one character at a time, after checking that it's not the first character of --> .

(?s) sets DOTALL mode (aka single-line mode), allowing the . to match newline characters.

If there's a possibility of two or more matches and you want to find them individually, you can replace the * with a non-greedy *? , like so:

"(?s)<!--(?:(?!-->).)*?<!DOCTYPE"

For example, applying that regex to the text of your question will find two matches, while the original regex will find one, longer match.

Answer 2

This seems like it is easily solved by using String.contains() :

if (yourHtml.contains("-->")) {
    // exclude
} else {
    // extract the content you need
    String content = 
        yourHtml.substring("<!--".length(), yourHtml.indexOf("<!DOCTYPE"));
}

I think you are looking too far into it.

Answer 3

\<!--([\s\S](?!--\>))*?(?=\<\!DOCTYPE)

this uses a negative lookahead to prevent the --> and a positive lookahead to find the <!DOCTYPE Here's a good reference for atomic assertions (lookahead and behind) .

Answer 4

I don't have a testing system handy so i can't give you the regex but you should look inside the Pattern documentation for something called negative lookahead assertion . This allows you to express rules of the form: Match this if not followed by that.

It should help you:)

Answer 5

A regular expression might not be the best answer to your problem. Have you tried splitting the first line away from everything else and seeing if it contains the --> ?

Specifically, something like:

String htmlString;
String firstLine = htmlString.split("\r?\n")[0];
if(firstLine.contains("-->"))
    ;//no match
//match