Re: Regexp::Approx - Use fuzzy regular expressions

It seems that your "fuzzy" regex isn't much different from just using dot and the length of the target. That is:

#!/usr/bin/perl -w
use strict;
use lib '.';
use Regexp::Approx 'fuzzy_qr';
use re 'eval';

while(<DATA>) {
    chomp;
    print "Using '$_' as fuzzy component:\n";
    my $fuzzy_part = fuzzy_qr( $_ );
    my $rx = qr/((?:$fuzzy_part\s)?\w+\d)$/;
    if ( "5678 DELAWARE AVENUE AOT 123" =~ $rx ) {
        print "\tRX Match: $1\n";
    }
    my $really_fuzzy = '.{' . length($_) . '}';
    my $rfx = qr/((?:$really_fuzzy\s)?\w+\d)$/;
    if ( "5678 DELAWARE AVENUE AOT 123" =~ $rfx ) {
        print "\tRFX Match: $1\n";
    }
}
__END__
APT
XXX
^%@#!(
fuzzy matches anything?
x

# output:
Using 'APT' as fuzzy component:
        RX Match: AOT 123
        RFX Match: AOT 123
Using 'XXX' as fuzzy component:
        RX Match: AOT 123
        RFX Match: AOT 123
Using '^%@#!(' as fuzzy component:
        RX Match: UE AOT 123
        RFX Match: UE AOT 123
Using 'fuzzy matches anything?' as fuzzy component:
        RX Match: 678 DELAWARE AVENUE AOT 123
        RFX Match: 678 DELAWARE AVENUE AOT 123
Using 'x' as fuzzy component:
        RX Match: T 123
        RFX Match: T 123
[download]

Comment on Re: Regexp::Approx - Use fuzzy regular expressions Download Code

Replies are listed 'Best First'.
Re: Re: Regexp::Approx - Use fuzzy regular expressions by diotalevi (Canon) on Nov 12, 2003 at 14:32 UTC
Drat. I musta screwed up the failure case while working on the send/recv functions. It did work correctly before. ~~Fixes to come in an hour or so.~~ Got it. The code was testing the return value of print() instead of the actual match response. When I added the general function wrapper I forgot to put a `return _read_msg()` in. (Again) it works correctly for me using your sample script.	[reply]
Re: Re: Re: Regexp::Approx - Use fuzzy regular expressions by Anonymous Monk on Nov 12, 2003 at 15:36 UTC
There seems to be remaining problems, for example: #!/usr/bin/perl -w use strict; use lib '.'; use Regexp::Approx 'fuzzy_qr'; use re 'eval'; while(<DATA>) { chomp; print "Using '$_' as fuzzy component:\n"; my $fuzzy_part = fuzzy_qr( $_ ); my $rx = qr/($fuzzy_part)/; if ( "5678 DELAWARE AVENUE AOT 123" =~ /$rx/ ) { print "\tRX Match: $1\n"; } } __END__ APT A XXX ^%@#!( fuzzy matches anything? x #### output: Using 'APT' as fuzzy component: RX Match: AO Using 'A' as fuzzy component: RX Match: 5 Using 'XXX' as fuzzy component: amatch: $_ is undefined: what are you matching? Compilation failed in require at xxx.pl line 4, <GEN0> line 51. BEGIN failed--compilation aborted at xxx.pl line 4, <GEN0> line 51. Using '^%@#!(' as fuzzy component: Broken pipe ### platform: $ perl -v This is perl, v5.8.0 built for i686-linux [download]	[reply] [d/l]


No such thing as a small change
	PerlMonks