This command removes empty lines:
sed -e '/^$/d' file
But, how do I remove spaces from beginning and end of each non-empty line?
Asked
Active
Viewed 2.5k times
4 Answers
41
$ sed 's/^ *//; s/ *$//; /^$/d' file.txt
`s/^ *//` => left trim
`s/ *$//` => right trim
`/^$/d` => remove empty line
kev
- 146,428
- 41
- 264
- 265
-
4`sed -r 's/^\s*//; s/\s*$//; /^$/d'` – kev Dec 19 '11 at 14:56
-
`sed -r "/^\s*\$/d ; s@^\s*(.*)\s*\$@\1@"` - less `sed`-"calling" :) – uzsolt Dec 19 '11 at 18:36
-
1Note to OS X users: the equivalent answer is `sed -E 's/^ +//; s/ +$//; /^$/d' file.txt` (the equivalent to the generalized, `\s`-based variant is: `sed -E 's/^[[:space:]]+//; s/[[:space:]]+$//; /^$/d' file.txt`. (There is no `-r` option for `sed` on OS X (as of 10.8); `-E` turns on extended regular expressions, which support `[[:..:]]`-style character classes, but apparently not their shorthand equivalents such as `\s`.) – mklement0 Aug 28 '12 at 02:47
-
This can be slightly improved with the accepted answer here: http://stackoverflow.com/questions/16414410/delete-empty-lines-using-sed/24957725#24957725 – ConMan Jul 25 '14 at 14:11
-
rather than space it might be better to use [[:blank:]] which includes also tabs – Petr Kozelka Dec 13 '14 at 02:06
8
Even more simple method using awk.
awk 'NF { $1=$1; print }' file
NF selects non-blank lines, and $1=$1 trims leading and trailing spaces (with the side effect of squeezing sequences of spaces in the middle of the line).
oguz ismail
- 39,105
- 12
- 41
- 62
M.S. Arun
- 773
- 10
- 21
2
Similar, but using ex editor:
ex -s +"g/^$/de" +"%s/^\s\+//e" +"%s/\s\+$//e" -cwq foo.txt
For multiple files:
ex -s +'bufdo!g/^$/de' +'bufdo!%s/^\s\+//e' +'bufdo!%s/\s\+$//e' -cxa *.txt
To replace recursively, you can use a new globbing option (e.g. **/*.txt).
kenorb
- 137,499
- 74
- 643
- 694