CODEWITHSUNDEEP
regexlinuxfilesed
bee
bee

Reputation: 1623

sed: Replace FIRST occurence of space with newline

I am trying to replace the first occurrence of a space in each line with a newline.

For instance

123 there is a monkey
567 there is also a tiger

would become

123
there is a monkey
567
there is also a tiger

I have tried doing this with the command:

sed -e 's/^\s\+/\n/g' output.txt > new.txt

with no luck. I am also open to any other programs in linux that would allow me to do this such as awk, perl, or bash.

Upvotes: 3

Views: 1776

Answers (5)

SGD
SGD

Reputation: 1716

Actually 

sed -e 's/\s/\n/' output.txt > new.txt

should suffice. If you drop the global flag it will match the first char and sed works by line.

Upvotes: 0

Mark Reed
Mark Reed

Reputation: 95242

Lots of answers that work in GNU sed but not BSD sed (such as you'll find on a Mac). BSD sed, for example, won't turn \n into a literal newline. Assuming you're in bash or modern ksh, you can use ANSI strings $'...' to get around that problem. All of the answers below work in either version.

First, if you know for sure it's just one space, you can do this:

sed $'s/ /\\\n/'

If you want to handle the case where there may be more than one space in a row, you can do this:

sed $'s/  */\\\n/'   # that's space, space, asterisk

Or you can use extended regular expressions:

sed -E $'s/ +/\\\n/'

If you want to handle arbitrary whitespace (e.g. tabs as well as spaces), then you can replace the literal space characters above with [[:space:]].

Alternately, you can use Perl, which has no version incompatibilities:

perl -pe 's/\s+/\n/'

Upvotes: 0

Wintermute
Wintermute

Reputation: 44023

Don't anchor the regex at the beginning of the line and ditch the g flag:

sed -e 's/\s\+/\n/' output.txt > new.txt

EDIT: BSD sed part removed because of pain.

Upvotes: 2

Jotne
Jotne

Reputation: 41436

You could use this:

awk '{sub(FS,RS)}1' file
123
there is a monkey
567
there is also a tiger

It changes the first field separator with record selector.

Upvotes: 0

anubhava
anubhava

Reputation: 784878

You can use:

sed $'s/ \{1,\}/\\\n/' file
123
there is a monkey
567
there is also a tiger

This works both on BSD as well on gnu sed.

Upvotes: 3

Related Questions