For large files (230K lines is at the small end of large) slurping files is a bad idea. In this case reading the second file first and building a lookup hash, then reading the first file and testing against the hash is probably the way to go. Consider:

use strict;
use warnings;

my $file1Str = <<END_FILE1;
2000/01/03/aaa/xyz.xml
2002/04/01/bbb/abc.xml
END_FILE1

my $file2Str = <<END_FILE2;
2000/01/03/xyz
END_FILE2

my %lookup;

open my $file2In, '<', \$file2Str or die ("Could not open file!");
while (my $line = <$file2In>) {
    my @parts = split '/', $line;
    my $key = join '/', @parts[0 .. 3];

    ++$lookup{$key};
}
close $file2In;

open my $file1In, '<', \$file1Str or die ("Could not open file!");
while (my $line = <$file1In>) {
    my @parts = split '/', $line;

    $parts[-1] =~ s/\.xml//i;

    my $key = join '/', @parts[0 .. 2, 4];

    next if exists $lookup{$key};
    print $line;
}
close $file1In;
[download]

Prints:

2002/04/01/bbb/abc.xml
[download]

Update Fixed wording - thanks jethro