How can I extract 3 or more separate text from a line using 'sed'
I have the following line:
echo <MX><[Mike/DOB-029/Post-555/Male]><MX>
So far I am able to extract the 'DOB-029' by doing
sed -n 's/.*\(DOB-[0-9]*\).*/\1/p'
but I am not getting the other texts such as the name or the post.
My expected output should be Mike DOB-029 Post-555
Edited
Say I have a list within a file and I want to extract specific text/IDs from the entire list and save it to a .txt file
sed 's/.*[\(.*\).\(DOB-[0-9]*\).\(Post-[0-9]*\).*/\1 \2 \3/' should do the trick!
Parts in between \( and \) are captured strings that can be called upon using \i with i the index of the group.
Script for custom use:
#! /bin/bash
fields=${1:-123}
file='/path/to/input'
name=$(sed 's/.*\[\([^\/]*\)\/.*/\1/' $file)
dob=$(sed 's/.*\(DOB-[0-9]*\).*/\1/' $file)
post=$(sed 's/.*\(Post-[0-9]*\).*/\1/' $file)
[[ $fields =~ .*1.* ]] && output=$name
[[ $fields =~ .*2.* ]] && output="$output $dob"
[[ $fields =~ .*3.* ]] && output="$output $post"
echo $output
Set the file with the line you want to parse in the file variable (I can add more functionality such as supplying the file as argument or getting it from a larger file if you like). And execute the script with an int argument, if this int contains '1' it will display name, if 2, it will display DOB and 3 will output post information. You can combine to e.g. '123' or '32' or whichever combination you like.
Stdin
If you want to read from stdin, use following script:
#! /usr/bin/env bash
line=$(cat /dev/stdin)
fields=${1:-123}
name=$(echo $line | sed 's/.*\[\([^\/]*\)\/.*/\1/')
dob=$(echo $line | sed 's/.*\(DOB-[0-9]*\).*/\1/')
post=$(echo $line | sed 's/.*\(Post-[0-9]*\).*/\1/')
[[ $fields =~ .*1.* ]] && output=$name
[[ $fields =~ .*2.* ]] && output="$output $dob"
[[ $fields =~ .*3.* ]] && output="$output $post"
echo $output
Example usage:
$ chmod +x script.sh
$ echo '<MX><[Mike/DOB-029/Post-555/Male]><MX>' | ./script.sh 123
Mike DOB-029 Post-555
$ echo '<MX><[Mike/DOB-029/Post-555/Male]><MX>' | ./script.sh 12
Mike DOB-029
$ echo '<MX><[Mike/DOB-029/Post-555/Male]><MX>' | ./script.sh 32
DOB-029 Post-555
$ echo '<MX><[Mike/DOB-029/Post-555/Male]><MX>' | ./script.sh
Mike DOB-029 Post-555
A solution with awk:
echo "<MX><[Mike/DOB-029/Post-555/Male]><MX>" | awk -F[/[] '{print $2, $3, $4}'
We set the delimiter as / or [ (-F[/[]). then we just print the fields $2, $3 and $4 which are the 2nd, 3rd and 4th fields respectively.
With sed:
echo "<MX><[Mike/DOB-029/Post-555/Male]><MX>" | sed 's/\(^.*\[\)\(.*\)\(\/[^/]*$\)/\2/; s/\// /g'
sed, I'd have suggested awk too. A scripting language with the power of a programming language, there's just no way to beat it :)
use the bash substitution builtins.
line="<MX><[Mike/D0B-029/Post-555/Male]><MX>";
linel=${line/*[/}; liner=${linel%\/*}; echo ${liner//\// }
/Male? What is the logic underlying the problem? Extract everything/separated?