sed command to collapse whitespace

__John_L. · 02-20-2021, 02:50 PM

I'm trying to write a sed command to collapse white space (tab, newlines, spaces, etc.) and whittle consecutive characters down to one space.

I tried:

Code:

sed "s/\s+/ /gm" infile.txt >outfile.txt

and also:

Code:

sed "s/\s{1,}/ /gm" infile.txt >outfile.txt

but both have no effect on whitespace in the input. So much for walking with sed.

Thanks in advance for any help you can provide.

shruggy · 02-20-2021, 02:57 PM

+ is a meta-character in extended regular expressions only. But sed uses basic regular expressions by default.

Code:

sed 's/\s\+/ /gm' infile.txt >outfile.txt

or

Code:

sed -r 's/\s+/ /gm' infile.txt >outfile.txt

or

Code:

sed -E 's/\s+/ /gm' infile.txt >outfile.txt

But they all won't match across the lines despite the m flag. For this, I'd rather

Code:

tr -s '[:space:]' ' ' <infile.txt >outfile.txt

syg00 · 02-20-2021, 05:30 PM

Be careful to define exactly what you want. No indented lines you want to retain - paragraph or lists for example ?. Newlines ?. Why keep a space at eol ?.

MadeInGermany · 02-21-2021, 01:22 AM

Standard/Posix sed (portable):

Code:

sed "s/[[:space:]]\{1,\}/ /g"