CODEWITHSUNDEEP
bashsedgrep
askaban
askaban

Reputation: 11

Extract only a part of data from a file

My input is test.txt which contains data in this format:

'X'=>'ABCDEF',
'X'=>'XYZ',
'X'=>'GHIJKLMN',

I want to get something like:

'ABCDEF',
'XYZ',
'GHIJKLMN',

How do I go about this in bash?

Thanks!

Upvotes: 1

Views: 56

Answers (4)

sjsam
sjsam

Reputation: 21955

Using awk

awk 'BEGIN{FS="=>"}{print $2}' file
'ABCDEF',
'XYZ',
'GHIJKLMN',

FS in awk stands for field separator. The code inside BEGIN is executed only at the beginning, ie, before processing the first record. $2 prints the second field.

A more idiomatic way of putting the above stuff would be

awk 'BEGIN{FS="=>"}$2' file
'ABCDEF',
'XYZ',
'GHIJKLMN',

The default action in awk is to print the record. Here we explicitly mention what to print. ie $2.

Upvotes: 0

Mark
Mark

Reputation: 4455

Here's a solution using sed: 

curl -sL https://git.io/fjeX4 | sed 's/^.*>//'

Sed is passed a single command: s///. is a regex that matches any characters (.*) from the beginning of the line (^) to the last '>'. The is an empty string, so essentially sed is just deleting all the characters on the line up to the last >. As with the other solutions, this solution assumes that there is only one '>' on the line.

Upvotes: 1

miku
miku

Reputation: 187994

If the data is really uniform, then you could just run cut (on example input):

$ curl -sL https://git.io/fjeX4 | cut -d '>' -f 2
'ABCDEF',
'XYZ',
'GHIJKLMN',

You can see flag explanations on explainshell.

With awk, it would look similar:

$ curl -sL https://git.io/fjeX4 | awk -F '>' '{ print $2 }'
'ABCDEF',
'XYZ',
'GHIJKLMN',

Upvotes: 0

choroba
choroba

Reputation: 241748

If the input never contains the character > elsewhere than in the "fat arrow", you can use cut:

cut -f2 -d\> file
  • -d specifies the delimiter, here > (backslash needed to prevent the shell from interpreting it as the redirection operator)
  • -f specifies which field to extract

Upvotes: 1

Related Questions