[SOLVED] Matching pattern and extraction

Issue

I have a log file which looks like below. It shows that there are many file are missing. I would like to list it.

$cat datafile.txt
    /data/kay/20091012.csv
    gawk: cmd. line:4: (FILENAME=/data/kay/out/501_20091013.txt FNR=1) fatal: file not file
    /data/kay/20091014.csv
    gawk: cmd. line:4: (FILENAME=/data/kay/out/501_20091015.txt FNR=1) fatal: file not file
    gawk: cmd. line:4: (FILENAME=/data/kay/out/501_20091016.txt FNR=1) fatal: file not file
    /data/kay/20091017.csv
    gawk: cmd. line:4: (FILENAME=/data/kay/out/501_20091018.txt FNR=1) fatal: file not file

I would to list the dates on which the file is missing. My script is below:

$script.sh
awk '{if($1 -eq "gawk")print $4}' datafile.txt
echo ${echo $(awk '{if($1 -eq "gawk")print $4}' datafile.txt):-14,8}

It is showing error. "Bad substitution"

My desire output:

$outfile.txt
20091013
20091015
20091016
20091018

Solution

Using sed

$ sed -En '/^ +gawk/s/[^)]*_([^.]*).*/\1/woutput.txt' input_file
$ cat output.txt
20091013
20091015
20091016
20091018

Using awk

$ awk -F"[_.]" '/gawk/{print $3 > "output.txt" }' input_file
$ cat output.txt
20091013
20091015
20091016
20091018

Answered By - sseLtaH

Answer Checked By - Marilyn (WPSolving Volunteer)