Thanks for the reference. I've looked through it, and there's some good stuff there to digest. However, one of the examples in perlman:perldsc is almost identical to something I already tried, following an example in Programming Perl.

    for ($i = 0 ; $i <= $#diary_data ; $i++ ) {
        print "$i is { ";
        for $temp_keys ( keys %{ $diary_data[$i] } ) {
            print "$temp_keys=$diary_data[$i]{$temp_keys} ";
        }
        print "}\n";
    }
[download]

This still generates an error. I broke it down to do one step at a time, like this:

 for ($i = 0 ; $i <= $#diary_data ; $i++ ) {
        print "I: $i\n";
        for $temp_keys ( keys %{ $diary_data[$i] } ) {
            print "Key: $temp_keys\n";
            print "Data: $diary_data[$i]{$temp_keys}\n";
        }
    }
[download]

This gives

I: 0
Key: timestamp
Use of uninitialized value at ./send_referral.pl line 197.
Data: 
Key: value
Argument "NWORLASTE800 LOGIN FAILED\nNWORLASTE801 LOGIN FAIL\n\nWA..."
+ isn't numeric in helem at ./send_referral.pl line 197.
Bad index while coercing array into hash at ./send_referral.pl line 19
+7.
$
[download]

However, if I comment out the line that starts with print "Data...", it prints the correct keys for this hash:

I: 0
Key: timestamp
Key: value
Key: user
[download]

What confuses me is the fact that I'm following examples straight out of Perl references and getting these errors. I'm starting to wonder if I've found a bug in ARS.pm.