Java PatternSyntaxException: Unmatched closing '('

Question

I need to remove all the URLs found in Twitter messages. I have a file with around 200,000 such messages so speed is crucial! To do this I use Java as a programming language, here is an example of my code:

public String performStrip(){

    String tweet = this.getRawTweet();
    String urlPattern = "((https?|http)://(bit\\.ly|t\\.co|lnkd\\.in|tcrn\\.ch)\\S*)\\b";

    Pattern p = Pattern.compile(urlPattern,Pattern.CASE_INSENSITIVE);
    Matcher m = p.matcher(tweet);

    int i = 0;

    while (m.find()) {
        tweet = tweet.replaceAll(m.group(i),"").trim();
        i++;
    }

    return tweet;
}

This works fine in following cases:

http://t.co/nhWp9hldEH        -> (empty string)
http://t.co/nhWp9hldEH"       -> "
http://t.co/nhWp9hldEH)aaa"   -> aaa"
aaa(http://t.co/nhWp9hldEH"   -> aaa("
aaa(http://t.co/nhWp9hldEH)"  -> aaa()"

However, when I get to a case as follows:

http://t.co/nhWp9hldEH)aaa"

I get an error

java.util.regex.PatternSyntaxException: Unmatched closing ')' near index 21

http://t.co/nhWp9hldEH)aa

at java.util.regex.Pattern.error(Pattern.java:1924)
at java.util.regex.Pattern.compile(Pattern.java:1669)
at java.util.regex.Pattern.<init>(Pattern.java:1337)
at java.util.regex.Pattern.compile(Pattern.java:1022)
at java.lang.String.replaceAll(String.java:2210)
at com.anturo.preprocess.url.UrlStripper.performStrip(UrlStripper.java:47)
at com.anturo.preprocess.testing.ReadIn.<init>(ReadIn.java:35)
at com.anturo.preprocess.testing.Main.main(Main.java:6)

I already looked into multiple similar questions regarding this error, however none have worked so far... Hoping someone can help me out here.

Answer 1

The problem is that you may have regex special characters in a URL, as you can see.

Short solution: use Pattern.quote(). Your code would then be:

tweet = tweet.replaceAll(Pattern.quote(m.group(i)),"").trim();

Note: only available since JDK 1.5, but you do use this or better, right?

Another solution is to simply use .replace():

tweet = tweet.replace(m.group(i), "").trim();

Unlike what its name suggests with regards to .replaceAll(), .replace() does replace all occurrences; it is simply that it doesn't take a regex as a replacement string. See also .replaceFirst().

Last but not least, you seem to be misusing .group()! Your loop should be:

while (m.find())
    tweet = tweet.replace(m.group(), "").trim();

No need for the i variable here; m.group(i) will, for one match, return what is matched by capturing group i in your regex.