How to extract usernames out of Tweets?

Question

I have the following example tweet:

RT @user1: who are @thing and @user2?

I only want to have user1, thing and user2.

What regular expression can I use to extract those three names?

PS: A username must only contain letters, numbers and underscores.

Answer 1

This should do it (I used named captures for convenience):

.+?@(?[a-zA-Z0-9_]+):[^@]+?@(?[^\s]+)[^@]+?@(?[a-zA-Z0-9_]+)

Answer 2

Tested:

/@([a-z0-9_]+)/i

---

In Ruby (irb):

>> "RT @user1: who are @thing and @user2?".scan(/@([a-z0-9_]+)/i)
=> [["user1"], ["thing"], ["user2"]]

In Python:

>>> import re
>>> re.findall("@([a-z0-9_]+)", "RT @user1: who are @thing and @user2?", re.I)
['user1', 'thing', 'user2']

In PHP:

<?PHP
$matches = array();
preg_match_all(
    "/@([a-z0-9_]+)/i",
    "RT @user1: who are @thing and @user2?",
    $matches);

print_r($matches[1]);
?>

Array
(
    [0] => user1
    [1] => thing
    [2] => user2
)

Answer 3

Is a good idea include twitter text library [1] in your project to resolve this text issues.

twttr.txt.extractMentions("a very generic twitt with some @mention");

[1] https://github.com/twitter/twitter-text-js

Answer 4

/(?<!\w)@(\w+)/

The above covers the following scenario, which other answers in this thread do not:

• An @ sign that is not supposed to be a username, e.g. "my email is [e-mail]"
• Still allows a username that is at the beginning of a string, e.g. "@username lorem ipsum..."

Answer 5

try an iterator (findall) with this regex:

(@[\w-]+)

bye