Sed to delete blank spaces

Question

Does anyone know how to use Sed to delete all blank spaces in a text file? I haven been trying to use the "d" delete command to do so, but can't seem to figure it out

Volker Siegel · Answer 1 · 2016-11-15T23:24:40.163

68

What kind of "space"?

To "delete all blank spaces" can mean one of different things:

delete all occurrences of the space character, code 0x20.
delete all horizontal space, including the horizontal tab character, "\t"
delete all whitespace, including newline, "\n" and others

The right tool for the job

If sed it not a requirement for some hidden reason, better use the right tool for the job.

The command tr has the main use in translating (hence the name "tr") a list of characters to a list of other characters. As an corner case, it can translate to the empty list of characters; The option -d (--delete) will delete characters that appear in the list.

The list of characters can make use of character classes in the [:...:] syntax.

tr -d ' ' < input.txt > no-spaces.txt
tr -d '[:blank:]' < input.txt > no-spaces.txt
tr -d '[:space:]' < input.txt > no-spaces.txt

When insisting on `sed`

With sed, the [:...:] syntax for character classes needs to be combined with the syntax for character sets in regexps, [...], resulting in the somewhat confusing [[:...:]]:

sed 's/ //g' input.txt > no-spaces.txt
sed 's/[[:blank:]]//g' input.txt > no-spaces.txt
sed 's/[[:space:]]//g' input.txt > no-spaces.txt

edited Nov 15 '16 at 23:24

answered Oct 18 '14 at 23:56

Volker Siegel

13,065
5
49
65

1

+1, POSIX notation for a blank space is the appropriate way to go. – Avinash Raj Oct 19 '14 at 02:26
1

Great hint to use tr for this task, works great. Only thing I had to adjust was that tr reads input from stdin, so what worked for me was tr -d ' ' < input.txt > no-spaces.txt. – Sky Sep 04 '16 at 18:50
@Sky's comment works in my macOS (not tested in Linux). Maybe the answer should be updated? – iplus26 Nov 13 '16 at 08:16
Thank a lot @Sky, that's indeed a bad error. I'll fix the three lines with tr now. (Do you see it anywhere else?) – Volker Siegel Nov 15 '16 at 23:20
Note that even [:space:] does not strip all unicode spaces. – starbeamrainbowlabs Apr 29 '21 at 21:10

score 11 · Answer 2 · answered Oct 16 '14 at 20:28

11

You can use this to remove all whitespaces in file:

 sed -i "s/ //g" file

answered Oct 16 '14 at 20:28

Cyrus

5,554

3

No dice. I've tried a similar approach too. Doesn't work – Justin Oct 16 '14 at 20:30
Do you want to delete whitespaces or tabs? – Cyrus Oct 16 '14 at 20:33
1

delete the whitespace – Justin Oct 16 '14 at 20:37
2

A simple test: man sed | sed "s/ //g" – Cyrus Oct 16 '14 at 21:14
I use sed "s: ::g" when i'm greping things, what's the difference between the two expression notations? – starscream_disco_party Apr 20 '17 at 01:13
1

@starscream_disco_party: it‘s the same. You can replace all 3/ with a different character, e.g.: sed "ss ssg" – Cyrus Apr 20 '17 at 05:17
You may want to try without the -i flag (--in-place) first — pipe | to less for instance. – user598527 Apr 07 '23 at 09:59

score 4 · Answer 3 · answered Apr 16 '20 at 05:34

4

perhaps this way too late, but sed takes regular expression as input. '\s' is the expression of all whitespaces. So sed s/'\s'//g should do the job. cheers

answered Apr 16 '20 at 05:34

aluk.e

41

Sed to delete blank spaces

3 Answers3

What kind of "space"?

The right tool for the job

When insisting on sed

When insisting on `sed`