CODEWITHSUNDEEP
linuxstringawk
Lin Ma
Lin Ma

Reputation: 10139

replace consecutive spaces with one comma using command line

Wondering the most efficient solution to replace one or more consecutive spaces with one comma? A single command is preferred. :)

I am using Mac OS/Linux.

thanks in advance, Lin

Upvotes: 1

Views: 125

Answers (4)

karakfa
karakfa

Reputation: 67467

This is what tr provides out of the box

tr -s ' ' ','

Upvotes: 2

Thomas Dickey
Thomas Dickey

Reputation: 54505

Noting the comment about leading/trailing spaces for this example (and suggesting that OP might not want to convert those):

sed 's/ \+/,/g' file

and that most people would consider TAB a "space", one could use a more comprehensive (but complicated) approach:

sed \
    -e ':loop' \
    -e 's/\([^[:space:]]\)[[:space:]][[:space:]]*\([^[:space:]]\)/\1,\2/g' \
    -e 't loop' file

The loop is needed to handle the case where the \1 and \2 overlap.

Here is an example input:

total 36
drwx------ 5 root root  4096 Aug 21 03:58 0994576031
drwx------ 4 tom  users 4096 Aug 21 04:27 fake-tom
-rw-r--r-- 1 tom  users    0 Aug 21 04:53 foo
-rw-r--r-- 1 tom  users  155 Aug 21 04:27 gpgagent.log
drwx------ 2 tom  users 4096 Aug 21 04:27 gpg-MOxBtc
drwx------ 2 root root  4096 Aug 21 03:58 kde-root
drwx------ 2 tom  users 4096 Aug 21 04:27 ssh-cdcgKy3228
drwxrwxrwt 2 root root  4096 Aug 21 03:57 VMwareDnD
drwxr-xr-x 2 root root  4096 Aug 21 03:58 vmware-root
drwx------ 2 root root  4096 Aug 21 03:58 vmware-root-2999462734

and output

total,36
drwx------,5,root,root,4096,Aug,21,03:58,0994576031
drwx------,4,tom,users,4096,Aug,21,04:27,fake-tom
-rw-r--r--,1,tom,users,0,Aug,21,04:53,foo
-rw-r--r--,1,tom,users,155,Aug,21,04:27,gpgagent.log
drwx------,2,tom,users,4096,Aug,21,04:27,gpg-MOxBtc
drwx------,2,root,root,4096,Aug,21,03:58,kde-root
drwx------,2,tom,users,4096,Aug,21,04:27,ssh-cdcgKy3228
drwxrwxrwt,2,root,root,4096,Aug,21,03:57,VMwareDnD
drwxr-xr-x,2,root,root,4096,Aug,21,03:58,vmware-root
drwx------,2,root,root,4096,Aug,21,03:58,vmware-root-2999462734

Upvotes: 0

John1024
John1024

Reputation: 113824

Using sed:

sed 's/  */,/g'

As an example:

$ echo 'a b  c   d' | sed 's/  */,/g'
a,b,c,d

Upvotes: 1

Kent
Kent

Reputation: 195039

give this one-liner a try, this will not replace the spaces at the BOL or EOL

awk -v OFS="," '{$1=$1}7' file

this line will replace all spaces:

awk '1+gsub(/ +/,",")' file

another one with sed, this will replace all spaces, including leading and ending ones

sed 's/ \+/,/g' file

Upvotes: 2

Related Questions