CODEWITHSUNDEEP
perl
ado
ado

Reputation: 1471

Counting how many times a combination of words appear in a given line using Perl

Dear Stackoverflow and Perl comrades:

I have a little question about Perl: I'm writing a log reader. The log format is like this

    2013-05-27T19:01:23 [INFO] item_id:1, state:start at Reader.pm line 23
    2013-05-27T19:01:29 [INFO] item_id:2, state:pause at Reader.pm line 23
    2013-05-27T19:01:30 [INFO] item_id:1, state:start at Reader.pm line 23

...

My goal is to count how many state:start a given item_id, for example, item_id:1, shows. In this case, it should be 2.

So far what I have thought about is a counter of words:

    sub count {
    my $count_start = 0;

    open (MYFILE, $file_location) or die "Wrong filename";
    while ($file_location = <MYFILE>){
            while ($file_location =~ /\bstart\b/ig){
                    $count_start++;
            }
    }
    close (MYFILE);
    return $count_start;
    }

But I have to count not how many times "start" appears but how many times "start" and "id" appear in the same line. I know I have to add some regular expression but cant figure out anything. Any ideas?

Regards!

Upvotes: 1

Views: 154

Answers (2)

ikegami
ikegami

Reputation: 385645

Assuming the order of the field is predictable:

my %counts;
while (<>) {
    ++$counts{$1} if /item_id:(\S+), state:start/;
}

Upvotes: 2

mogul
mogul

Reputation: 4553

A nasty one-liner for you:

mogul@guldager:~/tmp$ perl -MData::Dumper  -ne '$h{$1}++ if /(\d+), state:start/;END{print Dumper \%h}'<input-data.txt 
$VAR1 = {
          '1' => 2
        };

Upvotes: -1

Related Questions