newbie wants to count nbr of substr in a text string. How? 2

[JCHallgren]

I am VERY new to this! And have looked thru what i have avail here...not sure what I'm looking for...

I have a text string (comments from a guestbook) that i want to see how many times certain strings occur in it..
The key one is "

http://"

as more than 2 will be a sign of a spam post...is there an EASY way to scan $usrcoments for various strings and get a count of how many matches I get?

My book mentions Grep, but not clue on how to use it...
I would like to ideally match if mixed case used also...
And having all other similar words in a list to use ("poker","drug",etc) would be GREAT!

THANKS!

 

[missbarbell]

This should get you started:

my $text = "guestbook comment";
my @list = ('[URL unfurl="true"]http://','poker',[/URL] 'drug');
foreach my $substr (@list) {
  my $cnt = $text =~ /($substr)/g;
  print "text contains $cnt occurrances of '$substr'\n";
}

Using a scalar to collect regex matches, returns the number of matches, rather than the list of matches.

Barbie
Leader of Birmingham Perl Mongers

http://birmingham.pm.org

 

[stevexff]

Although you might be able to use map or grep, the following is easier to understand if you are new to perl.

use strict;
use warnings;

my @spamwords = ("poker", "[URL unfurl="true"]http://",[/URL] "drug", "enlargement", "viagra", "ad nauseum");
my $usercomments = "Make big money playing Internet Poker at [URL unfurl="true"]http://www.scumbag.com.[/URL] Genuine testimonial: 'I used to have a drug problem, but now I can afford it' (cannot be named for legal reasons), Essex";
my $spamindex = 0;

foreach my $word (@spamwords) {
    $spamindex += $usercomments =~ s/$word/$word/gi;
}

print "Spam rating is $spamindex: Bin it!\n" if $spamindex > 1;

The global case-insensitive search and replace s/$word/$word/gi returns the number of replacements made.

 

[stevexff]

OK, so you don't need to replace it to get the count.

$spamindex += $usercomments =~ /$word/gi;

works just as well... (thanks Barbie).

 

[ishnid]

That's some quality creative spamming there Steve. You've worked in the industry haven't you?

 

[stevexff]

No, but I used to be an AOL customer...

 

[missbarbell]

No that's just an beginners attempt, you need far more random and transposed (eg 'v1agr4') strings in there ;-)

I work in the industry, but on the side of the good guys, at MessageLabs.

Barbie
Leader of Birmingham Perl Mongers

http://birmingham.pm.org

 

[JCHallgren]

Stevexff and Barbie: That was just PERFECT! I like code that I can understand! And it seems to solve ALL the issues I have!!!
So far, all the junk has been using "normal" words...just a BUNCH of them so thats why the count would be good...one or two words may be normal chat, but 25 are not!

 

[JCHallgren]

Update: I tried it in my lil test pgm...doesn't seem to work right Only appears to count the FIRST occurance of each bad word...

 

[mlibeson]

This works:

$| = 1;

$text = <<EOT;

Hello testing counting hello string.

Just to make sure count is correct, hEllo!

Hello, Hello, Hello!!!
EOT

$text2 = "Hello";

$cnt = $text =~ s/(\bhello\b)/$1/gi;

print STDOUT "$cnt\n";

$cnt += $text2 =~ s/(\bhello\b)/$1/gi;

print STDOUT "$cnt\n";

OUTPUT:

6
7

Matching does not return how many times it matches the string, but search and replace does.



Michael Libeson

 

[missbarbell]

In that case you may need to do the following:

my $text = $usercomments;
$spamindex += $text =~ s/$word//gi;

I've discovered that a global match doesn't work the way I expected. I have made a copy of the variable $usercomments in the event you want to keep the original.

Barbie
Leader of Birmingham Perl Mongers

http://birmingham.pm.org

 

[JCHallgren]

Found via a IRC chat that adding in a " = () = " after $spamindex also works...thanks M Libeson also!

 

[cntr]

Awk
[tt]
BEGIN { FS=","
spamwords="poker,

http://,drug,enlargement,viagra,ad

nauseum"
usercomments="Make big money playing Internet Poker at

http://www.scumbag.com.

Genuine testimonial: 'I used to have a drug problem, but now I can afford drugs' (cannot be named for legal reasons), Essex"
$0 = spamwords
for (i=1; i<=NF; i++)
print $i, split(tolower(usercomments),junk,$i)-1
}
[/tt]

 

[missbarbell]

$spamindex += () = $usercomments =~ /($substr)/g;

Apparently the match only returns true or false. However, forcing into list context will return the right results.

Barbie
Leader of Birmingham Perl Mongers

http://birmingham.pm.org

 

[JCHallgren]

Barbie, don't worry...the guy on PerlMonks I got aid from at 3am, gave same initial answer...at that time, I had not noticed a reply here...so tried there for a 2nd opinion..

But now we BOTH know the way to do so!

 

[stevexff]

So that means I was right the first time, with the search & replace. I feel much better now...