topic Re: PERL pattern matching in Operating System - Linux

PERL pattern matching

Pat Tom — Sat, 28 Jul 2007 13:42:22 GMT

I have a flat file which contains hundreds of patterns. I need to write a PERL script that parses this file and returns each unique pattern and the count of its occurance.

I am trying it for quite sometime but not able to code it.

Can anyone please help.

Thanks
Pat

Re: PERL pattern matching

H.Merijn Brand (procura — Sat, 28 Jul 2007 14:48:07 GMT

#!/opt/perl/bin/perl

use strict;
use warnings;

my %pat;
while (<>) {
$pat{$_}++;
}
foreach my $pat (sort keys %pat) {
printf "%6d %s", $pat{$pat}, $pat;
}

Enjoy, Have FUN! H.Merijn

Re: PERL pattern matching

Dennis Handly — Sat, 28 Jul 2007 21:01:57 GMT

What do you mean by patterns? A file can have tokens or words but RE patterns is something only a human can find. After all, one pattern can be made for all English words.

If you want to search for space delimited alphabetic tokens you can use:
$ tr -cs "[A-Z][a-z]" "[\012*]" < file | sort | uniq -c

Re: PERL pattern matching

James R. Ferguson — Sun, 29 Jul 2007 15:30:31 GMT

Hi Pat:

Merijn's script gives you the solution based on the contents of a *line*. If your lines contain multiple "words" this variation gives you their unique counts.

#!/usr/bin/perl
use strict;
use warnings;

my %pat;
while (<>) {
my @a=m{\w+}g;
$pat{$_}++ for (@a);
}
foreach my $pat (sort keys %pat) {
printf "%6d %s\n", $pat{$pat}, $pat;
}

The \w regular expression matches an alphanumeric character or underscore but not a hyphen, quote, comma, semicolon, colons, etc.

As for writing the word "PERL" -- don't -- we are speaking of the "Perl" language:

http://www.perl.org/about/style-guide.html

You have previously posted another question about pattern matching for which solutions were provided. You forgot to score those solutions. It would be appreciated:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1100015

Regards!

...JRF...