Using sed to replace string in file by using regex capture group

Question

I start by saying I am very new on shell script so please don't shoot me !! :)

I have a file that contains the following text:

1   :    / 
2   :    /string-1/ 
4   :    /string-2/ 
5   :    /string-3/

and I like to remove the end slashes of the strings so the result should be like that :

1   :    / 
2   :    string-1 
4   :    string-2 
5   :    string-3

I have try that by using the sed as following:

local_export_folder="/home/merianos/Documents/repositories/DatabaseBackup/tmp"

sed -i -e 's/\/([^\/]+)\//\1/' ${local_export_folder}/blogs.csv

but this doesn't work.

Am I doing anything wrong ?

Your title is misleading, there are no 'regular expression variables' here. Can you be more specific than 'this doesn't work'? — Tom Fenech
– Tom Fenech, Commented Dec 29, 2014 at 14:14
What I try to describe be the title is how to use the content of the parentheses as the replacement text. I don't know how to write that title better. — KodeFor.Me
– KodeFor.Me, Commented Dec 29, 2014 at 14:15
The part between the parentheses is called a capture group. Hope that helps. — Tom Fenech
– Tom Fenech, Commented Dec 29, 2014 at 14:17

Tom Fenech · Accepted Answer · 2014-12-29 14:33:52Z

4

This works for me:

sed 's#/\([^/]*\)/#\1#' file

This captures any content between two forward slashes. The replacement is the content, without the slashes.

One issue that you may have been facing is that + (one or more) isn't understood by all versions of sed. I have changed it to a * (which means zero or more), which is more widely recognised. If you prefer, you could use the POSIX-compliant \{1,\} to mean one or more instead.

Output:

$ sed 's#/\([^/]*\)/#\1#' file
1   :    /
2   :    string-1
4   :    string-2
5   :    string-3

Depending on your version of sed, you may be able to use the -r or -E switch to enable extended regular expressions. This means that parentheses don't need escaping to be used as capturing groups and the + is understood:

sed -r 's#/([^/]+)/#\1#' file

edited Dec 29, 2014 at 14:33

answered Dec 29, 2014 at 14:24

Tom Fenech

75.1k13 gold badges119 silver badges154 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

KodeFor.Me Over a year ago

Thanks, while I used my regex, finally you helped me a lot, because I see that you escape the capture group parentheses too. So, when I escaped the parenthesis worked for me too.

unigeek Over a year ago

And in case this is helpful, to destructively change the input, use: sed -i 's#/([^/]*)/#\1#' file

Tom Fenech Over a year ago

@Merianos I have mentioned a way of using the parentheses without having to escape them.

Collectives™ on Stack Overflow

Using sed to replace string in file by using regex capture group

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related