CODEWITHSUNDEEP
replacesed
brucezepplin
brucezepplin

Reputation: 9752

sed - replace all strings that begin with

Hi I want to find all strings anywhere in a file that begin with the letters rs i.e.

rs12345 100
rs54321 200
300 rs13579

and delete all strings that begin with the criteria so that I get:

100
200
300

i.e. replace the string with nothing. I am not bothered about leading whitespace before final output as I deal with that later. I have tried sed 's/rs*//g' however this gives:

12345 100
54321 200

i.e. only removes the rs.

How do I edit my sed command to delete the entire string? Thanks

Upvotes: 3

Views: 20185

Answers (2)

fedorqui
fedorqui

Reputation: 289495

You can replace from starting rs up to a space by an empty string, so that the rsXXX gets removed.

sed 's/^rs[^ ]*//' file

This supposed your rs was in the beginning of the line. If rs can appear anywhere in the file, use:

sed 's/\brs[^ ]*//' file

The \b works as word boundary, so that things like hellorshello does not match.

Test

$ cat a
rs12345 100
rs54321 200
hellooo 300
$ sed 's/^rs[^ ]*//' a
 100
 200
hellooo 300

Note I am not dealing with the whitespace, since you mention you are handling it later on. In case you needed it, you can say sed 's/^rs[^ ]* \+//' file.

rs anywhere:

$ cat a
rs12345 100
rs54321 200
hellooo 300
and rs12345 100
thisrsisatest 24
$ sed 's/\brs[^ ]*//' a
 100
 200
hellooo 300
and  100
thisrsisatest 24

Note you current approach wasn't working because with rs* you are saying: r followed by 0 or more s.

Upvotes: 9

pacholik
pacholik

Reputation: 8972

Let the word be a pattern starting with word break \b and having characters \w (and ending with \b too, but we will not need it).

The command that removes words starting with rs will be sed -r 's/\brs\w+//g' file

$ cat file
rs12345 100
rs54321 200
ars1234 000
$ sed -r 's/\brs\w+//g' file
 100
 200
ars1234 000

Upvotes: 0

Related Questions