RegEx: words with two letters repeated twice (eg. ABpoiuyAB, xnvXYlsdjsdXYmsd)

Question

I had two regex tasks to do today -- and I did one properly and failed with the other. the first task was to find -- in a long, long text -- all the words beginning with "F" and ending with a vowel:

(\bf)\w*([euioay]\b)

and it worked perfectly.

the second one is way too difficult for a philology student ;-) I have to find all the words with repeated at least twice two-letter sequences, for example:

• tatarak is TATArak, "TA" twice;
• brzozowski is brZOZOwski, "ZO" twice;
• loremipsrecdks is loREmipsREcdks, "RE" twice;

can I have some help please? thanks in advance ;-)

Answer 1

Let's see:

(\w{2}) matches two letters (or digits/underscore, but let's ignore that) and captures them in group number 1. Then \1 matches whatever was matched by that group. So

\b\w*(\w{2})\w*\1

is what you're looking for (you don't need {2,} because if three letters are repeated, two letters are also repeated. Not checking for more than two makes the regex much more efficient. You can stop matching after the \1 backreference has succeeded).

Answer 2

This pattern ought to do the trick

\b\w*?(\w{2})\w*?\1\w*?\b

• \b is a word boundry
• \w*? some number of letters (lazily)
• (w{2}) exactly two letters, match and capture
• \w*? same as above
• \1 the content of our two letter capture group
• \w*? same as above
• \b another word boundry

A quick test in java:

public static void main(String[] args) {
   final Pattern pattern = Pattern.compile("\\b\\w*?(\\w{2})\\w*?\\1\\w*?\\b");
   final String string = "tatarak brzozowski loremipsrecdks a word that does not match";
   final Matcher matcher = pattern.matcher(string);
   while(matcher.find()) {
       System.out.println("Found group " + matcher.group(1) + " in word " + matcher.group());
   }
}

Output

Found group ta in word tatarak
Found group zo in word brzozowski
Found group re in word loremipsrecdks