[SOLVED] Need to insert a newline after 3rd line of grep pattern search

zabi · 12-18-2013, 01:35 AM

I have a file e.g

Fedora
Linux
MAC
Windows

it should grep 3 lines
e.g grep -A2 Fedora file

The o/p is

Fedora
Linux
MAC

And I want to insert a new line after 3rd line of last grep so o/p should be:

Fedora
Linux
MAC
NEWLINE
Windows

How to do it with[sed,awk,grep]??

berndbausch · 12-18-2013, 02:13 AM

Quote:

Originally Posted by zabi

I want to insert a new line after 3rd line of last grep so o/p should be:

Fedora
Linux
MAC
NEWLINE
Windows

How to do it with[sed,awk,grep]??

This awk script does it:

Code:

/Fedora/        { F=1 }
F>3             { printf "\n"; F=0 }
F               { F++ }
                { print }

When the line contains the string Fedora, a variable F is initialised to 1. We use this variable to count the lines after "Fedora".
When F is greater than 3, it's time to print a newline. Our job is done for now; reset F.
When F is non-zero, increment it.
And finally, print the input line.

I am sure it can be done in a much more elegant way.

grail · 12-18-2013, 02:34 AM

Code:

awk '/Fedora/{x=NR+3}NR==x{print ""}1' file

zabi · 12-18-2013, 03:04 AM

Perfect it worked!!
Thanks!!

zabi · 12-18-2013, 03:08 AM

i want the file to be edited at the same time,Currently it looks like this:

[XXXXXX # awk '/Fedora/{x=NR+3}NR==x{print "DON"}1' sdf
Fedora
Linux
MAC
DON
Windows
[XXXXXX # cat sdf
Fedora
Linux
MAC
Windows
[XXXXX #

berndbausch · 12-18-2013, 03:26 AM

Code:

 
# cp sdf sdf.backup-in-case-something-goes-wrong
# awk '/Fedora/{x=NR+3}NR==x{print "DON"}1' sdf >sdf.temp
# mv sdf.temp sdf

zabi · 12-18-2013, 03:53 AM

So there is no other way like sed -i works....
Anyways... Thanks a lot!

druuna · 12-18-2013, 04:47 AM

Using sed:

Code:

$ cat input
Fedora
Linux
MAC
Windows
$ sed -i.bak '/Fedora/ {N;N;s/$/\nDON/}' input
$ cat input
Fedora
Linux
MAC
DON
Windows

the .bak following the -i option makes a backup, just in case.

zabi · 12-18-2013, 05:06 AM

Great!!
Thanks a lot!!

zabi · 12-18-2013, 11:01 PM

resolved

danielbmartin · 12-24-2013, 11:04 AM

Quote:

Originally Posted by berndbausch

Code:

 
awk '/Fedora/{x=NR+3}NR==x{print "DON"}1' sdf >sdf.temp

With this InFile ...

Code:

Fedora
Linux
MAC
Windows
Fedora
Derby
Stetson
Bowler
... et cetera ...

... this code ...

Code:

awk '/Fedora/{x=NR+3}NR==x{print "DON"}1' $InFile >$OutFile

... produced this OutFile ...

Code:

Fedora
Linux
MAC
DON
Windows
Fedora
Derby
Stetson
DON
Bowler
... et cetera ...

Now, let's allow the line to be inserted to be defined externally.
With the same InFile, this code ...

Code:

new="DON"
awk -v new=$new '/Fedora/{x=NR+3}NR==x{print new}1' $InFile >$OutFile

... produced the same OutFile. So far, so good.

Now, let's change the line to be inserted, making it contain blanks.
With the same InFile, this code ...

Code:

new="D O N"
awk -v new=$new '/Fedora/{x=NR+3}NR==x{print new}1' $InFile >$OutFile

... produced an error message ...

Code:

awk: cmd. line:1: fatal: cannot open file `N' for reading (No such file or directory)

How should this be coded?

Daniel B. Martin

druuna · 12-24-2013, 11:32 AM

@danielbmartin: You need to use quotes in the awk variable assignment part:

Code:

$ new="D O N"
$ awk -v new="$new" '/Fedora/{x=NR+3}NR==x{print new}1' infile 
Fedora
Linux
MAC
D O N
Windows
Fedora
Derby
Stetson
D O N
Bowler

danielbmartin · 12-24-2013, 12:39 PM

Quote:

Originally Posted by druuna

@danielbmartin: You need to use quotes in the awk variable assignment part ...

That does it. Thank you!

Daniel B. Martin