Removing lines with only a single comma

Question

I have search for a way to do this, I have found similar questions but nothing that matches this.

I have a cvs file with the following text:

nfoabc,infidel
infoghi,infojkl
,
,
,
inferno,infopqr
infoabc,infodef
,
,
,

The desired output is:

infoabc,infidel
infoghi,infojkl
inferno,infopqr
infoabc,infodef

I'm wondering if there is a way to filter the lines with a single , and no other characters… I think this is where I'm having difficulty.

Any assistance would be greatly appreciated. TIA.

Possible duplicate of [How to skip lines matching a string](http://stackoverflow.com/questions/6684857/how-to-skip-lines-matching-a-string) — Benjamin W., Feb 11 '16 at 18:21
Thanks everyone. What I was doing wrong was not writing it back to a file, ideally, I'd have it overwrite the existing file but I can deal with this. Thanks! — cducasse, Feb 11 '16 at 18:42
@cducasse: `cmd file > tmp && mv tmp file`. Replace `cmd` with whatever command you want to execute. — Ed Morton, Feb 11 '16 at 18:50

score 3 · Answer 1 · answered Feb 11 '16 at 18:28

3

$ grep -v '^,$' file
infoabc,infidel
infoghi,infojkl
inferno,infopqr
infoabc,infodef

or for this particular input file where there's either 1 or 0 commas:

$ grep '[^,]' file
infoabc,infidel
infoghi,infojkl
inferno,infopqr
infoabc,infodef

answered Feb 11 '16 at 18:28

Ed Morton

score 1 · Answer 2 · answered Feb 11 '16 at 18:53

1

If you want to edit the file, you can always use ed:

ed -s file <<< $'g/^,$/d\nw'

answered Feb 11 '16 at 18:53

gniourf_gniourf

Cool! Can you do it with MS-DOS `edlin` too? ;-) – Mark Setchell Feb 11 '16 at 19:39
@MarkSetchell I have no idea what `edlin` is, sorry `:(`. – gniourf_gniourf Feb 11 '16 at 19:42

score 0 · Answer 3 · answered Feb 11 '16 at 18:35

0

awk to the rescue!

$ awk '!/^,$/' file

nfoabc,infidel
infoghi,infojkl
inferno,infopqr
infoabc,infodef

answered Feb 11 '16 at 18:35

karakfa

3 Answers3