Unfortunately, this is a difficult problem... because what you want is fuzzy matching, and regex engines don't make it easy to do fuzzy matching.

You could try a different approach, for example, split the string into appropriate segments and then calculate some similarity measure to the target words, and accept/reject based on some threshold. There are some modules out there that might assist you with that, e.g. String::Similarity or Text::Levenshtein.

In addition to almut's suggestions you might look at String::Approx. I've used it in the distant past and liked it. I have no idea if it's still as good as the other options for what you're trying to do.

How to check in a line words "publish" and "date" exists or not. I have to print the "matched" if publish and date both exists in line

while(<DATA>){
$ =~ m/publish|date/;
}
__DATA__
The book was published on the date "20-08-2009".
[download]

If you want an exact match, that's easy:

while (<DATA>){
    print "matched\n" if /publish/ and /date/;
}
[download]

But note that this would also match "the date was published on"...

If the order in which the words must occur is relevant, you could do

while (<DATA>){
    print "matched\n" if /publish.*date/;
}
[download]

How to match any of the strings below.

Match any:

#!/usr/bin/perl
while (<DATA>){
    chomp;
    if ($. == 1) {
        print "'$_' matched\n";
    }
}
[download]

Match all:

#!/usr/bin/perl
while (<DATA>){
    chomp;
    if (1) {
        print "'$_' matched\n";
    }
}
[download]

You'll probably have to be more specific. Like by saying what you want to not match.

How can I match all the typos

Start by making a list of all the typos?

You might also be interested in finding the distance between the string being tested against a reference string. That's the name of insertions, changes and deletions needed to turn one of the strings into the other. The lower the distance, the more similar they are.

Terrible spec: title and body content conflict. - -, and then change, repeatedly, in later nodes.

Working from the title, however,

#!/usr/bin/perl
use strict;
use warnings;
# 794700

while(<DATA>){
        # if($_ =~ m/\bpublication date\b/){
    if($_ =~ m/\bpublication date\s/){
    print $_;
    }
}


__DATA__
These are the dates.
The publication date is 20-10-09
The publicatiion dateee is 20-05-08
The publication dat is 10-06-07
The publication date's are 20-10-09,29-02-08
[download]

Output with \b after "date" (incorrectly accepts "date's" which should be "dates" in this --plural-- context.)

perl 794700.pl
The publication date is 20-10-09
The publication date's are 20-10-09,29-02-08
[download]

Output with \s after "date"

perl 794700.pl
The publication date is 20-10-09
[download]

If one wishes to accept both "date" and "dates" (eg:

The publication date is 20-10-09
The publication dates are 20-10-09, 29-02-08)
[download]

alternation could be an answer:

Please see perldoc perlre (perlre) and company.

use strict;
use Text::Aspell;

my $suggest = ( @ARGV and $ARGV[0] eq '-s' ) ?  shift : '';

my $speller = Text::Aspell->new or die "dang it\n";

while (<>) {
    for my $word ( grep /^[a-z]+$/i, split ) {
        next if ( $speller->check( $word ));  # skip if word is in dic
+tionary
        print "misspelled word: $word\n";
        if ( $suggest ) {
            my @suggestions = $speller->suggest( $word );
            my $advice = ( @suggestions )
                ? join( " ", "Might be one of:", @suggestions )
                : "No idea what that should be.";
            print "  $advice\n\n";
        }
    }
}
[download]