OK, if anyone else stumbles upon this page, here's what I came up with...Apparently, $datasource is a reference to an array. So you can't just change the $ to an @. You have to use curly braces like this to read into the reference:

$element = scalar(@{$datasource});
[download]

My code now reads:

  1 #!/usr/bin/env perl
  2 
  3 use strict;
  4 use warnings;
  5 #use IO::Socket;
  6 use Data::Dumper;
  7 use XML::Simple;
  8 my $xs1 = XML::Simple->new();
  9 my $file = $ARGV[0];
 10 my $doc = $xs1->XMLin($file, forcearray => 1);
 11 foreach my $datasource ($doc->{DATASOURCE}){
 12   my $length = scalar(@{$datasource});
 13   print $length ."\n";
 14   foreach my $element (@{$datasource}){
 15     print Dumper($element);
 16 #  my $element = scalar(@datasource);
 17 #  print "Element = $element\n"; 
 18 #  for(my $element=0; $element < scalar($datasource); $element++){
 19 #      print $datasource[$element];
 20   }
 21 }
[download]

Each $element is a reference to a hash now. I'll have to de-reference those to access them as well, I'm sure.

In most languages, you can only use an identifier in one context - not so in Perl! You can have $x (scalar), @x (array), %x (hash), sub x{...} all in the same program. So $x is a completely unrelated thing to @x. As you have correctly surmised, the value of an array evaluated in a scalar context is the length of that array.

$element = scalar(@{$datasource});
$element = @$datasource; #same thing $element is a scalar so explicit conversion not required.
$x < @$datasource #also scalar context
You only need the extra curly braces when there is a subscript involved - so deference applies to the whole thing.