looking for lines... 1

[ogniemi]

hallo,

input file (cron log file with lines) - mixed lines like:
root : CMD ( /home/den/proc400.ksh >/dev/null ) : PID ( 18940110 ) : Tue Nov 6 21:30:00 2012
root : CMD ( /home/den/clean.ksh full >/dev/null ) : PID ( 53412070 ) : Tue Nov 6 23:30:00 2012
Cron Job with pid: 18940110 Failed
Cron Job with pid: 18943456 Successful
Cron Job with pid: 19535560 Successful
Cron Job with pid: 53412070 Failed

and want to find all tasks which exited with error (Failed status).

What i wanted to do is:

for i in `grep "^Cron Job with pid" log|awk '{print $5}'|sort -u`
do
awk -vx=$a '$(NF-7) == x || $(NF-1) == x {print}' log|sed 'N;s/\n/,/'|grep Failed$
done

and then join every two lines and grep for Failed.

The problem is that
$(NF-7) == x || $(NF-1) == x
is doesn't do what I expect. I want it to find all lines having the PID number ($a) on position $(NF-1) or $(NF-7). is it possible knowing that some lines have less fields then 7?

 

[feherke]

Hi

Possible to misunderstood the requirement, as you posted no sample of expected output. This is based on the "join every two lines and grep for Failed" part :

awk '[blue]NF[/blue][teal]>[/teal][purple]10[/purple][teal]&&[/teal][navy]$(NF-9)[/navy][teal]==[/teal][green][i]"PID"[/i][/green][teal]{[/teal]p[teal][[/teal][navy]$(NF-7)[/navy][teal]]=[/teal][navy]$0[/navy][teal];[/teal][COLOR=chocolate]next[/color][teal]}[/teal][navy]$NF[/navy][teal]==[/teal][green][i]"Failed"[/i][/green][teal]{[/teal][COLOR=chocolate]print[/color] p[teal][[/teal][navy]$(NF-1)[/navy][teal]],[/teal][navy]$0[/navy][teal]}[/teal]' /input/file

Tested with [tt]gawk[/tt] and [tt]mawk[/tt].

Feherke.
[link feherke.github.com/]

http://feherke.github.com/

[/url]

 

[ogniemi]

you simply got it right ;-)

thank you!!!