Reverse search query

[ranjit]

[tt]
Could someone please assist with the following problem.
Search file for the string ##EVENT### and print the 2 previous lines, as per the desired output:

Input file:
AAA-BBB-C-CC1FIN7438
AAA-BBB-C-CC2FIN222831
AAA-BBB-C-CC3FIN222832
##EVENT###
AAA-BBB-CCC1PROD48408
AAA-BBB-CCC3PROD4843
AAA-BBB-CCC4PROD4084
##EVENT###
AAA-BBB-CCC1TEC803083202
AAA-BBB-CCC2TEC4893
AAA-BBB-CCC3TEC325
AAA-BBB-CCC4TEC4336
##EVENT###
BLANK-----#
BLANK-----#
##EVENT###

Desired ouptut:
AAA-BBB-C-CC2FIN222831
AAA-BBB-C-CC3FIN222832
AAA-BBB-CCC3PROD4843
AAA-BBB-CCC4PROD4084
AAA-BBB-CCC3TEC325
AAA-BBB-CCC4TEC4336
BLANK-----#
BLANK-----#
[/tt]

 

[PHV]

And what have you tried so far ?

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[feherke]

Hi

awk '/##EVENT###/{print s1;print s2}{s1=s2;s2=$0}' /input/file
[gray]# or[/gray]
grep -B 2 "##EVENT### /input/file | grep -v "##EVENT###"

Feherke.

http://rootshell.be/~feherke/

 

[ranjit]

[tt]
Thanks for the solution. My approach was rather clumsy...that is, reversing the file order (sending the output to a temp file) and then using something like:
$0 ~ /##EVENT###/ {
getline
{
for(i=1;i<=2;i++) {
printf("%s\n", $0)
getline
}
}
}

However it didn't work in all cases.

[/tt]

 

[feherke]

Hi

Nice try, ranjit. Your only mistake was using the [tt]getline[/tt] one time more then needed :

/##EVENT###/ {
  for(i=1;i<=2;i++) {
    getline
    print
  }
}

And do not call your code clumsy. It is a good start. Ok, maybe is not the optimal one. But there are uglier solutions too, for example in [tt]sed[/tt] :

sed -n '/##EVENT###/{g;p};x;s/.*\n//;x;H' /input/file

Feherke.

http://rootshell.be/~feherke/

 

[ranjit]

Many thanks Feherke for the assistance.