use Data::Dump qw/dump/;
use DBI;

my $dbh = DBI->connect(
    "dbi:CSV:",
    "", "",
    {
	    #f_schema => 'test',
        f_dir    => '.',
        f_ext    => '.csv',
    }
) or die "Could not open database: $DBI::errstr";
$dbh->{RaiseError} = 1;


$dbh->{csv_tables}->{file1} = {
	file  => 'vm.csv',
};

$dbh->{csv_tables}->{file2} = {
	file  => 'vfiler_ds.csv',
};

my $query = "SELECT file1.vm_name, vm_cluster, vFiler_IP, vFiler_cluster FROM file1 join file2 WHERE file1.vm_name = file2.vm_name";
my $result = $dbh->selectall_arrayref($query);

print dump($result);

Overly complicated. Why use file and file1? That doesn't say anything in the queries, making them hard to read.

my $dbh = DBI->connect ("dbi:CSV:", undef, undef, {
    f_dir      => ".",
    f_ext      => ".csv/r",
    RaiseError => 1,
    PrintError => 1,
    }) or die "Could not open database: $DBI::errstr";

my $result = $dbh->selectall_arrayref (join " " =>
    "SELECT vm.vm_name, vm_cluster, vFiler_IP, vFiler_cluster",
    "FROM   vm JOIN vfiler_ds",
    "WHERE  vm.vm_name = vfiler_ds.vm_name");
[download]

Using f_ext - which you already do - just add /r to make the extension required, and just use the file names as table names in the query.

As the f_*** attributes are just DBI attributes (passed to DBD::File eventually), you can put RaiseError and other DBI attributes in the same hash.


Enjoy, Have FUN! H.Merijn