[SOLVED] Find the text and copy to previous line

sunilsagar · 02-13-2012, 03:09 AM

Hello All,

I have to find the word, if found copy the text to previous line.

Before :

ABC is working
DEF is working
GHI is not-working
ACVFGDSGDDSEDSSS is
working
XYZ is working

Now, if awk '{ print $1}' is working, then the output will be

ABC is working
DEF is working
GHI is not-working
ACVFGDSGDDSEDSSS is working
XYZ is working

grail · 02-13-2012, 04:52 AM

hmmmm ... not sure I can see what that awk statement has to do with anything as it will obviously not produce the output you have shown.

Based on the data you have provided, the following could work:

Code:

awk 'ORS=/working$/?"\n":"\0"' file

firstfire · 02-13-2012, 07:23 AM

Hi.

The same in sed:

Code:

sed  '/working$/!{N; s/\n/ /}' file

@grail: Wow, cool! I'd add a space before '\0':

Code:

awk 'ORS=/working$/?"\n":" \0"' file

vikas027 · 02-13-2012, 07:35 AM

Hi Sunil,

If I understood you need some thing like this.

Say, your present file is as below

Code:

root@box2:/tmp# cat /tmp/oldfile
one is working
two is working
three is working
string is working
four is working
five is working
six is working
string is working
root@box2:/tmp#

Now, you need your file as below, say you need to search line containing keyword "string"

Quote:

one is working
two is working
three is working
string is working
string is working
four is working
five is working
six is working
string is working
string is working

grail · 02-13-2012, 09:15 AM

Quote:

Originally Posted by firstfire

I'd add a space before '\0'

I would have but on copying the OPs data it already has one at the end of the previous line

sunilsagar · 02-13-2012, 10:12 PM

Yes Vikas

You found it right. I need similar scenario. Thanks

grail · 02-14-2012, 01:18 AM

Well I must say I have no idea how Vikas worked this out based on your explanation, but an easy solution is:

Code:

awk '/what your looking for/{print}1' file

sunilsagar · 02-14-2012, 02:00 AM

Actually I have a file with below content :

ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is
STARTED
ADMU0508I: The Application Server "Server3" is STOPPED

But when I execute the command "awk 'ORS=/STARTED$/?"\n":" \0"' 55.txt"

I am getting the output : Its pulling the line after the stop.

[root@hlixgr217 tmp]# awk 'ORS=/STARTED$/?"\n":" \0"' 55.txt
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Application Server "Server3" is STARTED

I need the output as :
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Application Server "Server3" is STARTED

Please advise. Thanks

firstfire · 02-14-2012, 02:26 AM

Hi.

I already posted this solution:

Code:

$ sed '/is$/{N;s/\n/ /}'  infile.txt
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Application Server "Server3" is STOPPED

It we encounter the word `is' at the end of line (`$'), then append next line to pattern space (`N'; Now pattern space looks like "...line ending with is\nnext line", \n -- newline character), then replace newline with space. Of cource this solution assumes that next line is a correct one (STARTED or STOPPED). We could check that too, though.

chidam · 02-14-2012, 02:33 AM

@Grail - Its working.
What is the meaning of that?

vikas027 · 02-14-2012, 04:33 AM

Quote:

Originally Posted by grail

Well I must say I have no idea how Vikas worked this out based on your explanation, but an easy solution is:

Code:

awk '/what your looking for/{print}1' file

Awesome Grail.

This is the reason, I love this forum. I would do this with some loops and sed/awk combinations. Really, in my wildest dreams I could not think that this could be a one liner.

Grail, request you to please explain how this worked.

vikas027 · 02-14-2012, 04:50 AM

Quote:

Originally Posted by sunilsagar

Actually I have a file with below content :

Code:

ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED 
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is
            STARTED
ADMU0508I: The Application Server "Server3" is STOPPED

But when I execute the command "awk 'ORS=/STARTED$/?"\n":" \0"' 55.txt"

I am getting the output : Its pulling the line after the stop.

Code:

[root@hlixgr217 tmp]# awk 'ORS=/STARTED$/?"\n":" \0"' 55.txt 
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is            STARTED
ADMU0508I: The Application Server "Server3" is  STARTED

I need the output as :

Code:

ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED 
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Application Server "Server3" is STARTED

Please advise. Thanks

Sunil, this is NOT something which you had asked earlier. Also, try to use CODES when pasting outputs of commands, it makes post more readable.

vikas027 · 02-14-2012, 06:42 AM

Hi Sunil,

This script should work for you.

Code:

#!/bin/bash
cp -v 55.txt 55.txt_bkup

updatefile ()
{
string=$1
grep -n ^$string 55.txt_bkup | awk -F":" '{print $1}' > lines

for line in `cat lines`;
do
line1=$(expr $line - 1)
sed -i ''$line1's/is/is '$string'/' 55.txt_bkup
done

grep -v ^$string 55.txt_bkup > 55.txt_bkup.tmp
cat 55.txt_bkup.tmp > 55.txt_bkup
rm -f 55.txt_bkup.tmp
}

updatefile STARTED
updatefile STOPPED

rm -f `pwd`/lines

I created a test file, 55.txt and run the script. The script will create a file 55.txt_bkup which should be your required file.

Code:

root@box2:/tmp/abc# cat 55.txt
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED 
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is
STARTED
ADMU0508I: The Application Server "Server3" is STOPPED
ADMU0508I: The Application Server "Server2" is
STARTED
ADMU0508I: The Web server "HTTPServer" is
STOPPED
ADMU0508I: The Web server "HTTPServer" is
STOPPED 
ADMU0508I: The Application Server "Server2" is
STARTED
root@box2:/tmp/abc#


root@box2:/tmp/abc# 
root@box2:/tmp/abc# ./abc.sh 
`55.txt' -> `55.txt_bkup'
root@box2:/tmp/abc# 

root@box2:/tmp/abc# cat 55.txt_bkup
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED 
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Application Server "Server3" is STOPPED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED
ADMU0508I: The Web server "HTTPServer" is STOPPED
ADMU0508I: The Application Server "Server2" is STARTED
root@box2:/tmp/abc#

I am sure there can be an easier way, may be a one liner to do this task.
But as per my limited knowledge in shell scripting, this should work for you.

grail · 02-14-2012, 07:24 AM

Quote:

Originally Posted by chidam

@Grail - Its working.
What is the meaning of that?

You might need to tell me where you got this line from chidam??

grail · 02-14-2012, 07:42 AM

OP - you will find that the awk statement does work, however, your new data now has an additional space at the end of the following line:

Quote:

ADMU0508I: The Web server "HTTPServer" is STOPPED $

If you need to be able to have white space at the end you need to advise so it can be adjusted.
Assuming the white space is not there, the following works just fine:

Code:

$ cat -A input_file
ADMU0508I: The Application Server "Portal" is STARTED$
ADMU0508I: The Web server "HTTPServer" is STOPPED$
ADMU0508I: The Node Agent "nodeagent" is STARTED$
ADMU0508I: The Application Server "Server2" is$
STARTED$
ADMU0508I: The Application Server "Server3" is STOPPED$
$ awk 'ORS=/ED$/?"\n":" \0"' input_file
ADMU0508I: The Application Server "Portal" is STARTED
ADMU0508I: The Web server "HTTPServer" is STOPPED
ADMU0508I: The Node Agent "nodeagent" is STARTED
ADMU0508I: The Application Server "Server2" is STARTED
ADMU0508I: The Application Server "Server3" is STOPPED

The cat I used with -A was to illustrate that there are no spaces before the end of any line.