CODEWITHSUNDEEP
awksed
Nathan
Nathan

Reputation: 107

sed with text and multiple delimiters

I am using sed to remove some text and replace it, I can get most of it accomplished but I'm at a stand still with replacing the string of 5 colon's before and the one after of it. Any thoughts on where I'm going wrong?

Text file I have:

The user name1 ONE:TWO:THREE:FOUR:FIVE:NewName:SIX
The user name2 SEVEN:EIGHT:NINE:TEN:ELEVEN:NewName:TWELVE
The user name3 Thirteen:Fourteen:Fifteen:Sixteen:NewName:Seventeen

command:

sed 's/^/ The user /'| sed 's/:.*/ My New Name is/'

Output:

The user name1 My New Name is

Expected output:

The user name1 My New Name is NewName

Upvotes: 1

Views: 900

Answers (3)

RavinderSingh13
RavinderSingh13

Reputation: 133508

With your shown samples please try following.

awk -F'[ :]' '{print $1,$2,$3,"My New Name is",$(NF-1)}' Input_file

Explanation: Simple explanation would be, setting field separator as space OR colon here for all lines. Then simply printing 1st, 2nd, 3rd fields followed by string May Name is followed by 2nd last field of current line as per requirements of OP.

Upvotes: 2

Wiktor Stribiżew
Wiktor Stribiżew

Reputation: 626802

You can use

sed '/^The user / s/[^: ]*:.*:\(.*\):.*/My New Name is \1/' file

See the online demo:

s='The user name1 ONE:TWO:THREE:FOUR:FIVE:NewName:SIX
The user name2 SEVEN:EIGHT:NINE:TEN:ELEVEN:NewName:TWELVE
The user name3 Thirteen:Fourteen:Fifteen:Sixteen:NewName:Seventeen'
sed '/^The user / s/[^: ]*:.*:\(.*\):.*/ My New Name is \1/' <<< "$s"
# => 
#    The user name1 My New Name is NewName
#    The user name2 My New Name is NewName
#    The user name3 My New Name is NewName

Here,

  • /^The user / - finds the lines that start with The user
  • s/[^: ]*:.*:\(.*\):.*/ My New Name is \1/ finds
    • [^: ]* - zero or more chars other than a colon and space
    • :.*: - :, any text, :
    • \(.*\) - Group 1 (\1): any text
    • :.* - : and the rest of the string.
  • \1 is the replacement placeholder, the contents of Group 1.

Upvotes: 1

anubhava
anubhava

Reputation: 785128

It is easier to use awk:

awk '{print $1, $2, $3, "My New Name is", a[split($NF, a, /:/)-1]}' file

The user name1 My New Name is NewName
The user name2 My New Name is NewName
The user name3 My New Name is NewName

Upvotes: 2

Related Questions