Rails - Apr 2006 - one-to-one relationship confusion

Hello,

I posted a similar email, earlier in the week, but I still must not be
getting it.

I have a customer that aggregates two addresses; a shipping address
and a billing address.  Addresses can also be referenced by other
objects in a similar way, so I would really like to avoid putting the
keys referencing the customer or other objects in the addresses table.
 Addresses are exclusively referenced by parent objects and a
many-to-many association is not appropriate.

Ideally I would like the customers table to use a shipping_address_id
and a billing_address_id to reference the aggregate addresses.  For
illustration purposes, another object, company might have a
main_address_id, an after_hours_address_id and a weekend_address_id.

In order to keep the keys in the customer table the belongs_to
statements are in the customer model and the has_one statement is in
the address model.

This creates the structure I am after, but not the behavior.  The
customer is really the parent object, when I delete the customer I
want it to cascade to addresses, etc.

It seems like I am writing more code than I should to propagate
updates, etc to addresses.  Am I totally off base?  Any guidance would
be greatly appreciated.

Thanks,
Dan

One solution might be to create a many-to-many relationship for each 
type of object that can have an address, but only ever put one address 
in the collection. So there would be an address_customer table with 
columns address_id, customer_id, and maybe address_type (indicating if 
it is billing or shipping). Then you would have tables such as 
address_supplier, address_employee, etc.

It''s a little more work up front, but the nice thing is that this would
be very flexible for future changes.

Mike Johnson


Dan Munk wrote:
> Hello,
>
> I posted a similar email, earlier in the week, but I still must not be
> getting it.
>
> I have a customer that aggregates two addresses; a shipping address
> and a billing address.  Addresses can also be referenced by other
> objects in a similar way, so I would really like to avoid putting the
> keys referencing the customer or other objects in the addresses table.
>  Addresses are exclusively referenced by parent objects and a
> many-to-many association is not appropriate.
>
> Ideally I would like the customers table to use a shipping_address_id
> and a billing_address_id to reference the aggregate addresses.  For
> illustration purposes, another object, company might have a
> main_address_id, an after_hours_address_id and a weekend_address_id.
>
> In order to keep the keys in the customer table the belongs_to
> statements are in the customer model and the has_one statement is in
> the address model.
>
> This creates the structure I am after, but not the behavior.  The
> customer is really the parent object, when I delete the customer I
> want it to cascade to addresses, etc.
>
> It seems like I am writing more code than I should to propagate
> updates, etc to addresses.  Am I totally off base?  Any guidance would
> be greatly appreciated.
>
> Thanks,
> Dan
> _______________________________________________
> Rails mailing list
> Rails@lists.rubyonrails.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>
>
>

Dan Munk wrote:
> This creates the structure I am after, but not the behavior.  The
> customer is really the parent object, when I delete the customer I
> want it to cascade to addresses, etc.
You could use a before_destroy callback in the customer for this.

class Customer < ActiveRecord::Base
   belongs_to :shipping_address...
   belongs_to :billing_address...
   before_destroy :clear_addresses
   private
     def clear_addresses
       self.shipping_address.destroy
       self.billing_address.destroy
     end
end

-- 
Alex