Is it possible to find sequences of words using regex?

Question

Is it possible to find sequences where the ending letter of one word is the same as the beginning letter of the next word, and the ending letter of that word is the same as the beginning letter of the next next word and so on?

For example:

elementum magna sodales should match elementum magna, while something like Proin nunc curna, aliquet nec should return Proin nunc curna, aliquet, but an earring should return nothing because n is not the same as e.

I've tried something like \w*(\w)[\s:;'",.?!]*\1\w* but that only matches two words, I kind of need them to daisy chain together.

Yes, it is possible. At least for any regex engine supporting backreferences. — Wiktor Stribiżew, Nov 24 '16 at 17:08
Are you asking this as a challenge (for fun), if not, what is your actual use case? There might be better ways to resolve this. — d4nyll, Nov 24 '16 at 17:23
@d4nyll yeah, for fun. maybe there is a better way, i don't know. — , Nov 24 '16 at 17:26
@CasimiretHippolyte I've tried `\w*(\w)[\s:;'",.?!]*\1\w*` but that only matches two words, and I kind of need to daisy chain it together. — , Nov 24 '16 at 17:29
You should put that in your question (edit it), the lack of attempts is probably the reason of the downvote. — Casimir et Hippolyte, Nov 24 '16 at 17:34

Casimir et Hippolyte · Answer 1 · 2016-11-24T17:48:55.740

4

You can do it with this pattern:

(?i)\b(?:[a-z]*([a-z])[^a-z]+(?=\1))+[a-z]*

Details:

(?i) # makes the pattern case-insensitive
\b
(?:  # non-capturing group: one word and eventual following non-word characters
    [a-z]*([a-z]) # a word with the capture of the last character
    [^a-z]+ # non-word characters
    (?=\1) # lookahead that checks the next word first letter
)+ # repeat
[a-z]* # last next word

demo

edited Nov 24 '16 at 17:48

answered Nov 24 '16 at 17:38

Casimir et Hippolyte

88,009
5
94
125

Wow, that was quick! Thank you very much! – Nov 24 '16 at 17:42

score 0 · Answer 2 · edited May 23 '17 at 10:32

0

Yes, is theoretically possible if your regex-engine supports recursive references.

This problems is similar with checking if a string is palindrome (question: How to check that a string is a palindrome using regular expressions?).

edited May 23 '17 at 10:32

Community

1
1

answered Nov 24 '16 at 17:15

Dario

3,905
2
13
27

1

You don't need to use recursion for this problem. (You can also find palindromes without recursion). – Casimir et Hippolyte Nov 24 '16 at 17:17

Is it possible to find sequences of words using regex?

2 Answers2