awk: How can I keep the format of the record?

quanba · 04-05-2010, 07:41 AM

Hi there,

In /etc/fstab, I have a record:

LABEL=/<\t\t\t>/<\t\t\t>ext3<\t\t\t>default,nolog<\t>1 1

\t: represent a tab between 2 fields

I just want to remove "nolog" in the 4th field (only): gsub(/nolog/,"",$4)

The function work ok but it returns a record:
LABEL=/ / ext3 default 1 1

I know the problem is OFS=" ", but how can I keep the format of the record? (the same number of tabs, spaces).

Thanks for your help

ghostdog74 · 04-05-2010, 07:49 AM

why don't you just gsub?

Code:

awk '{gsub(",nolog","")}1' file

quanba · 04-05-2010, 08:13 AM

Quote:

Originally Posted by ghostdog74

why don't you just gsub?

Code:

awk '{gsub(",nolog","")}1' file

I thought about your solution, but I want to gsub at the 4th field only, for example:

/tmp/nolog/abc / ext3,nolog

In this case, gsub will replace nolog in the directory name

grail · 04-05-2010, 09:12 PM

Firstly, your assumption about replacing nolog in the directory name is incorrect as there is no preceding comma (,).

But I am not sure why you feel the format has changed with your original example?
There appears to be the same number of white space in both the original and the removed version on the line.

Maybe you would clarify a little further?

quanba · 04-05-2010, 09:22 PM

Quote:

Originally Posted by grail

Firstly, your assumption about replacing nolog in the directory name is incorrect as there is no preceding comma (,).

But I am not sure why you feel the format has changed with your original example?
There appears to be the same number of white space in both the original and the removed version on the line.

Maybe you would clarify a little further?

Sorry for the inconvenience, I edited the thread above

grail · 04-05-2010, 09:50 PM

Well unless you think you will have a comma in a directory name or that option anywhere else on the line,
it would appear ghostdog's solution is the winner