0

I have a file that contains lines that look like this:

username1
username2
./2011-05-25-0.json.gz:{"repo":{"url":"https://api.github.dev/repos//","name":"/"},"type":"FollowEvent","public":true,"created_at":"2011-05-25","payload":{"target":{"gravatar_id":"5e17","id":144929,"repos":12,"followers":1,"login":"username3"}},"actor":{"gravatar_id":"dec","id":21,"url":"https://api.github.dev/users/user","avatar_url":"https://secure.gravatar.com/avatar/decc?d=http://github.dev%2Fimages%2Fgravatars%2Fgravatar-user-420.png","login":"username4"},"id":"14"} 
./2011-05-25-0.json.gz:{"repo":{"url":"https://api.github.dev/repos//","name":"/"},"type":"FollowEvent","public":true,"created_at":"2011-05-25","payload":{"target":{"gravatar_id":"f9ed","id":82,"repos":5,"followers":4,"login":"username5"}},"actor":{"gravatar_id":"decc93","id":21,"url":"https://api.github.dev/users/username","avatar_url":"https://secure.gravatar.com/avatar/de?d=http://github.dev.user-420.png","login":"username6"},"id":"147"}

I would like to obtain a file that contains:

username1
username2
username4
username6

That is, if the string "login": occurs on a line, then obtain the string contained in " immediately after the second occurrence of "login". Otherwise, keep the line untouched.

How can I do this with grep, or sed?

StatsSorceress
  • 3,019
  • 7
  • 41
  • 82

1 Answers1

1

You can try:

sed 's/.*{.*login":"\([^"]*\)".*/\1/g' text.txt

Or 

sed -i.bkup 's/.*{.*login":"\([^"]*\)"}.*/\1/g' text.txt

To overwrite the file with the new usernames and bkup the old one to text.txt.bkup

NOTE: This assumes the username does not contain the pattern described by the regex.

vintnes
  • 2,014
  • 7
  • 16
Perplexabot
  • 1,852
  • 3
  • 19
  • 22