Find the occurence of particular string in log file

Shell Programming and Scripting
1

I have a log file which looks like this:

[2012-12-04 13:54:15,744] <845185415165:STATUS:5/0:0:0:0:0|ghy59DI5zasldf87asdfamas8df9asd903tGUVSQx4GJVSQ==>

I have to extract DATE and number of times the keyword STATUS is shown on each date.

Input is :

[2012-12-04 13:54:15,744] <1354625655744:STATUS:5/0:0:0:0:0|ghy59DI5ztGUVSQx4GJVSQ==>

Output: in excel file, Column1(date) Column2(Number of occurences)
My excel file should also contain all the dates, if on any particular day, STATUS keyword is not found, it will show count as 0

Please help me ASAP. I am in very urgent need of this shell script.

2 
perl -e 'while(<>){ 
  $counts{$1} = 0 if (/\[(\d{4}-\d{2}-\d{2})/ ) && (! defined $counts{$1})
  $counts{$1}++ if /\[(\d{4}-\d{2}-\d{2}).+STATUS/;
}
for $date (sort keys %counts){
  print "$date , $counts{$date}\n";
} '  filename.log >status_count.csv
3

thanks for the very quick reply..

But, can you please help me in writing in normal bash script for the same

4

Why do you need bash specifically perl can be called within a bash script, it's just another utility, a very powerful one that you can use to write complex programs in, but also a command line utility.

5

skrynesaver...

how do i call a bash script from the perl??

6

Perl allows you to call a shell command in three ways, exec, system and with backticks or using the qx(), quote executable, operator

exec ("$ENV{HOME}/bin/my_script.sh --f - l -a - g -s");
system ("$ENV{HOME}/bin/my_script.sh -f - l -a - g -s");
$result = qx($ENV{HOME}/bin/my_script.sh -f - l -a - g -s);

exec hands control over to the called operation and the entire process dies with the called script, whereas system forks a new process and waits for it to complete. The qx operator is like system except it returns the output of the command which allows us to assign it.

1 Like
7

Actually, I have no idea how to write the script. This is actually my first time.
I wrote the code like this:

#!/usr/local/bin/perl

MAIN:
{

        while(<>)
        {
                $counts{$1} = 0 if (/\[(\d{4}-\d{2}-\d{2})/ ) && (! defined $cou
nts{$1})
                $counts{$1}++ if /\[(\d{4}-\d{2}-\d{2}).+STATUS/;
        }

        for $date (sort keys %counts)
        {
                print "$date , $counts{$date}\n";
        }
}

On running the code, I am getting following error:

bash-3.00$ perl test.pl license.log > status.log
Scalar found where operator expected at test.pl line 9, near ")
                $counts"
        (Missing operator before $counts?)
syntax error at test.pl line 9, near ")
                $counts"
syntax error at test.pl line 9, near "++ if"
syntax error at test.pl line 16, near "}"
Execution of test.pl aborted due to compilation errors.

Can you help me in this?

8 
#!/usr/local/bin/perl

use strict;
use warnings;

my %counts;
while(<>){
    $counts{$1} = 0 if (/\[(\d{4}-\d{2}-\d{2})/ ) && (! defined $counts{$1});
    $counts{$1}++ if /\[(\d{4}-\d{2}-\d{2}).+STATUS/;
}
for my $date (sort keys %counts){
    print "$date , $counts{$date}\n";
}

I missed a semi-colon :((

9

This is a way to do it in awk:

awk -F'[[ ]' '{d[$2]}/STATUS/{d[$2]++}END{for(i in d)print i,d?d:0}' OFS=, filename