Call me lazy or even simple, but I implement "SQL libraries" using a simple subclass of DBIx::Simple.

package SmallProject::Database;
use base 'DBIx::Simple';

sub connect {
    my $self = shift;
    $self->SUPER::connect('dbi:...', 'u', 'p', { RaiseError => 1});
}

sub foo {
    my $self = shift;
    my ($where, @binds) = $self->abstract->where({ @_ });
    $self
        ->query("SELECT * FROM foo WHERE $where LIMIT 1", @binds)
        ->hash;
}
# This could be simpler if SQL::Abstract supported LIMIT clauses:
# sub foo {
#     my $self = shift;
#     $self->select(foo => '*', { @_ }, \1);
# }

sub foos {
    my $self = shift;
    $self->select(foo => '*', { @_ })->hashes;
}

1;
[download]

Somewhere else:

my $db = SmallProject::Database->new;

my $foo = $db->foo(id => 15);

my @active_foos = $db->foos(active => 1);
[download]

I really like Class::DBI in many ways, but it is often too slow for what I need. Note by the way that one should never SELECT * and then not request a hash :)