Java Regex strange thing

Question

I have tried this reg. expression in order to retrieve an email address. As i have little experience with that, i would like to ask you if you know what's wrong with it, since it doubles one word:

regexp = "(\\w+)(\\(at\\))((\\w+\\.)+)([a-z]{2,3})";

Supposing i have an input "madrugada(at)yahoo.co.uk", it gives out as a result [email protected] .

pattern = Pattern.compile (regexp);
m = pattern.matcher (my_input);
while (m.find()) {
    for (int i=0; i<=m.groupCount(); i++)
         // it would give out: madrugada (at) yahoo co co uk
}

Thank you

John Haager · Accepted Answer · 2012-03-14 21:42:33Z

3

You have an extra set of parentheses in your regex. When you loop through the capture groups, both of the capture groups (one of which is inside the other) are returned, duplicating the output since they captured the same thing.

Try this

regexp = "(\\w+)(\\(at\\))(\\w+\\.)+([a-z]{2,3})";

Edit: An alternate RegEx that uses non-capturing groups seems like it would solve the problem.

regexp = "(\\w+)(\\(at\\))((?:\\w+\\.)+)([a-z]{2,3})";

edited Mar 14, 2012 at 21:42

answered Mar 14, 2012 at 20:24

John Haager

2,1151 gold badge17 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

Madrugada Over a year ago

now it outputs stg like: [email protected] ( so it only takes the last word before .uk )

user1181445 Over a year ago

I believe this one may also work: "\w+([-+.']\w+)*@\w+([-.]\w+)*\.\w+([-.]\w+)*"

John Haager Over a year ago

Given the RegEx in the original question, it seems it might be better to pull only the matching groups that you know you want (1, 2, 3, and 5) and skip that inner capture group (4) that duplicates the final part of the non TLD portion of the email address.

Madrugada Over a year ago

thank you John, this is what i've done before but i wanted something more ellegant:)

John Haager Over a year ago

@Madrugada I just had a flash of insight. If we convert the inner domain name capture group into a non-capturing group, that should solve the problem. Try this RegEx: "(\\w+)(\(at\))((?:\\w+\\.)+)([a-z]{2,3})";

Alex · Accepted Answer · 2012-03-14 21:52:19Z

1

import java.util.regex.*;
String a="madrugada(at)yahoo.co.in.ro.uk";
String regexp="(\\w+)(\\(at\\))(\\w+)((?:\\.\\w+)*)(\\.[a-z]{2,3})";
Pattern pattern = Pattern.compile (regexp);
Matcher m = pattern.matcher (a);
while (m.find()) {
    for (int i=0; i<=m.groupCount(); i++)
         println m.group(i);
}

produces following output:

madrugada(at)yahoo.co.in.ro.uk
madrugada
(at)
yahoo
.co.in.ro
.uk

EDIT:

Updated the above with a non capturing group. The reason that it did not work before is even though it matched multiple .\w+ patterns, the backreference was only to the last one. Also changed the non capturing group to * for accomodate madrugada(at)yahoo.uk

edited Mar 14, 2012 at 21:52

answered Mar 14, 2012 at 20:50

Alex

11.2k9 gold badges42 silver badges64 bronze badges

4 Comments

Madrugada Over a year ago

i am sorry, this is worse :( output: [email protected]

Alex Over a year ago

I've edited the regex in the post, I've made the first match exclude the . and the match words starts with . and it seems to have worked for me

Madrugada Over a year ago

it works indeed.. did you change the expression meantime, thanks a lot!

Madrugada Over a year ago

but for this it doesn't work: madrugada(at)yahoo.co.in.ro.uk . I don't get it...it seems correct :(

Matt · Accepted Answer · 2012-03-15 04:55:53Z

1

You also don't really want to include m.group(0), as it contains the whole segment that matched your overall RE.

for (int i=1;i<=m.groupCount();i++) {
  System.out.println(m.group(i));
}

answered Mar 15, 2012 at 4:55

Matt

11.8k2 gold badges28 silver badges35 bronze badges

Collectives™ on Stack Overflow

Java Regex strange thing

3 Answers 3

5 Comments

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

5 Comments

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related