if you're just interested in *EXACTLY* the same files, use MD5 (get it from
search.cpan.org)
even one BIT of difference (literally) will result in a different checksum
thorough and fast
hth
Jos Boumans
----- Original Message -----
From: "Maxim Berlin" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, August 07, 2001 1:35 PM
Subject: Re: Compare large files memory error
> Hello Mbedish,
>
> Tuesday, August 07, 2001, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
> Mac> Is there a better way to compare large files than this snippet,
> Mac> which runs out of memory if files > 30mb.
>
> Mac> It is also slow, about the same speed as comparing in a text editor!
>
> Mac> Thank you.
>
> Mac> __SNIP__
>
> Mac> @file1 = (<IN1>);
> Mac> @file2 = (<IN2>);
> first, it is a very bad practice to read entire file into memory. for
> example, if file size is bigger than free memory, you'll have
unpredictable
> result.
> second, the slowest operation in your example is file I/O. your files
> may have differences in second string, so you do not need to read rest of
> file.
>
>
> Mac> $are_equal = compare_arrays(\@file1, \@file2);
>
> Mac> if ($are_equal) {
>
> Mac> print "Files are IDENTICAL \n";
> Mac> }
>
> Mac> else
>
> Mac> {
>
> Mac> print "Files are DIFFERENT \n";
> Mac> }
>
> Mac> sub compare_arrays {
> Mac> my ($first, $second) = @_;
> Mac> #no warnings; # silence spurious -w undef complaints
> third, you compare your arrays two times
>
> first pass:
> Mac> return 0 unless @$first == @$second;
> second pass:
> Mac> for (my $i = 0; $i < @$first; $i++) {
> Mac> return 0 if $first->[$i] ne $second->[$i];
> Mac> }
>
> Mac> return 1;
> Mac> }
>
> Mac> __SNIP__
> so, i can recommend something like
> $retval=1;
> while( $a=<IN1> )
> {
> $b=<IN2>;
> if ($a ne $b)
> {
> $retval=0; last;
> }
> }
> close IN1; close IN2;
> return $retval;
>
> at the beginning, you can compare files size (perldoc -f stat).
>
>
> Best wishes,
> Maxim mailto:[EMAIL PROTECTED]
>
>
>
> --
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
>
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
