help me with string splitting using sed

Answer 1

Can you use awk?

As pointed out by wilhelmtell and Colin, replacing every space by a comma is not the right way because it also replaces the spaces in a string like NT AUTHORITY\NETWORK SERVICE

Use this instead

$ cat temp | awk 'BEGIN{FS = ","}{print $1","$8","$10}' | sed 's/"//g'
wmiprvse.exe,NT AUTHORITY\NETWORK SERVICE,N/A

---

[Old answer]

$ cat temp | awk 'BEGIN{FS = ","}{print $1, $8, $10}' | sed 's/"//g'|
  sed 's/ /,/g'

wmiprvse.exe,NT,AUTHORITY\NETWORK,SERVICE,N/A

1.

awk 'BEGIN{FS = ","}{print $1, $8, $10}'

print column number 1, 8 and 10

2.

sed 's/"//g'

remove "

3.

sed 's/ /,/g'

replace space by a ,

Answer 2

If you can have (escaped) quotes or commas in values then regular expressions are not the right tool for the job. You will need a designated state-machine for this. Your best bet is to write a small script with Ruby, Python, Perl or the likes.

Answer 3

I think you want to use awk for this:

[jkramer/sgi5k:~]# cat foo
"wmiprvse.exe","3596","Console","0","5,632 K","Running","NT AUTHORITY\NETWORK SERVICE","0:00:00","N/A"
[jkramer/sgi5k:~]# awk -F'","|"' '{print $2 "," $8 "," $10}' foo
wmiprvse.exe,NT AUTHORITY\NETWORK SERVICE,N/A

Answer 4

I think the problem is in the thousand-delimiter comma. I would use cut/paste to split the file:

"wmiprvse.exe","3596","Console","0","5,632 K","Running","NT AUTHORITY\NETWORK SERVICE","0:00:00","N/A"

cat myfile | cut -d'"' -f2,14,18 | paste -d','

It cuts fields 2,14 and 18 taking " as a separator and then pastes them together with comma as a separator

Answer 5

Get your *nix tools from GNU win32 here. They are more updated. The more appropriate tool for this job is gawk, not sed. If you look at that massive regular expression, you will understand what i mean.

gawk "{print $1,$7}" file

Note, windows cmd.exe doesn't like double quotes. So if you are printing inside gawk and needs double quotes, always escape them, eg

gawk "{print \"hello\" }" file