Rails - Jul 2006 - How can I make has_many prevent a delete that would lead to orphans?

e.g.

class Asset < ActiveRecord::Base
  validates_presence_of :asset_number, :make, :model, :location,
:name, :serial_number
  validates_numericality_of :asset_number
  validates_uniqueness_of :asset_number
  belongs_to :user
  belongs_to :location
  belongs_to :asset_type, :foreign_key => ''type_id''
end

class Location < ActiveRecord::Base
  validates_presence_of :name
  validates_uniqueness_of :name
  has_many :assets
end


If I create a Location, with name ''London'' and then create an
Asset
with that location it makes sense for me to be able to delete the
asset but not to delete the location.

If the location is deleted I have a situation where the asset now has
a foreign key pointing to a non-existant record. In other words it
become orphaned.

I know that there are ways (in
http://api.rubyonrails.com/classes/ActiveRecord/Associations/ClassMethods.html)
to make the foreign keys be set to null, or to delete all the orphans
(a bit like ON DELETE CASCADE), but what I want is to prevent the
delete of the location if it has assets which belong to it.

I was expecting some way to do this would be fairly easy and built
into AR but my hopes of that are fading... Hopefully it''s just that
I''ve missed something though.

If there is no way to do it that is built into AR how might I mod AR
or something to add that functionality? Presumably some kind of call
to location.assets.is_empty? to check if the delete is going to be
allowed?

-Rob

-- 
------------------------------------------------------
"On two occasions, I have been asked [by members of Parliament],
''Pray, Mr. Babbage, if you put into the machine wrong figures, will
the right answers come out?'' I am not able to rightly apprehend the
kind of confusion of ideas that could provoke such a question."
-- Charles Babbage (1791-1871)

"98.5% of DNA is considered to be junk DNA with no known purpose.
Maybe it''s XML tags." -- Anon

http://www.robhulme.com/
http://robhu.livejournal.com/

On 7/21/06, Robert Hulme <rob@robhulme.com> wrote:
>
> e.g.
>
> class Asset < ActiveRecord::Base
>   validates_presence_of :asset_number, :make, :model, :location,
> :name, :serial_number
>   validates_numericality_of :asset_number
>   validates_uniqueness_of :asset_number
>   belongs_to :user
>   belongs_to :location
>   belongs_to :asset_type, :foreign_key => ''type_id''
> end
>
> class Location < ActiveRecord::Base
>   validates_presence_of :name
>   validates_uniqueness_of :name
>   has_many :assets
> end
>
>
> If I create a Location, with name ''London'' and then
create an Asset
> with that location it makes sense for me to be able to delete the
> asset but not to delete the location.
>
> If the location is deleted I have a situation where the asset now has
> a foreign key pointing to a non-existant record. In other words it
> become orphaned.
>
> I know that there are ways (in
>
>
http://api.rubyonrails.com/classes/ActiveRecord/Associations/ClassMethods.html
> )
> to make the foreign keys be set to null, or to delete all the orphans
> (a bit like ON DELETE CASCADE), but what I want is to prevent the
> delete of the location if it has assets which belong to it.
>
> I was expecting some way to do this would be fairly easy and built
> into AR but my hopes of that are fading... Hopefully it''s just
that
> I''ve missed something though.
>
> If there is no way to do it that is built into AR how might I mod AR
> or something to add that functionality? Presumably some kind of call
> to location.assets.is_empty? to check if the delete is going to be
> allowed?
>
> -Rob
>
> --
> ------------------------------------------------------
> "On two occasions, I have been asked [by members of Parliament],
> ''Pray, Mr. Babbage, if you put into the machine wrong figures,
will
> the right answers come out?'' I am not able to rightly apprehend
the
> kind of confusion of ideas that could provoke such a question."
> -- Charles Babbage (1791-1871)
>
> "98.5% of DNA is considered to be junk DNA with no known purpose.
> Maybe it''s XML tags." -- Anon
>
> http://www.robhulme.com/
> http://robhu.livejournal.com/
> _______________________________________________
> Rails mailing list
> Rails@lists.rubyonrails.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>
I think a before_destroy on the Location should automatically stop the
destroy action if a false is returned.

before_destroy :check_assets

protected

def check_assets
  self.assets.count == 0
end
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://wrath.rubyonrails.org/pipermail/rails/attachments/20060720/e4dd2b90/attachment.html

> I think a before_destroy on the Location should automatically stop the
> destroy action if a false is returned.
>
> before_destroy :check_assets
>
> protected
>
> def check_assets
>   self.assets.count == 0
> end
Thanks - that does work, but it makes me wonder why there isn''t a
built in way of saying to behave like this whenever there is a
has_many relationship.

Wouldn''t that be a sensible default? Otherwise you end up with orphans
so when you come to say list the assets Rails blows up.

Is there a simple way to propogate an error up ''Some assets still use
this location'' or something?

-Rob

-- 
------------------------------------------------------
"On two occasions, I have been asked [by members of Parliament],
''Pray, Mr. Babbage, if you put into the machine wrong figures, will
the right answers come out?'' I am not able to rightly apprehend the
kind of confusion of ideas that could provoke such a question."
-- Charles Babbage (1791-1871)

"98.5% of DNA is considered to be junk DNA with no known purpose.
Maybe it''s XML tags." -- Anon

http://www.robhulme.com/
http://robhu.livejournal.com/

On 7/21/06, Robert Hulme <rob@robhulme.com> wrote:
>
> > I think a before_destroy on the Location should automatically stop the
> > destroy action if a false is returned.
> >
> > before_destroy :check_assets
> >
> > protected
> >
> > def check_assets
> >   self.assets.count == 0
> > end
> Thanks - that does work, but it makes me wonder why there isn''t a
> built in way of saying to behave like this whenever there is a
> has_many relationship.
>
> Wouldn''t that be a sensible default? Otherwise you end up with
orphans
> so when you come to say list the assets Rails blows up.
>
> Is there a simple way to propogate an error up ''Some assets still
use
> this location'' or something?
>
> -Rob
>
> --
> ------------------------------------------------------
> "On two occasions, I have been asked [by members of Parliament],
> ''Pray, Mr. Babbage, if you put into the machine wrong figures,
will
> the right answers come out?'' I am not able to rightly apprehend
the
> kind of confusion of ideas that could provoke such a question."
> -- Charles Babbage (1791-1871)
>
> "98.5% of DNA is considered to be junk DNA with no known purpose.
> Maybe it''s XML tags." -- Anon
>
> http://www.robhulme.com/
> http://robhu.livejournal.com/
> _______________________________________________
> Rails mailing list
> Rails@lists.rubyonrails.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>
This could be a good plugin / patch to write and submit to the core team.
It does seem like a good default behavior.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://wrath.rubyonrails.org/pipermail/rails/attachments/20060720/4010f2d8/attachment.html

On Thursday, July 20, 2006, at 4:15 PM, Robert Hulme
wrote:
>> I think a before_destroy on the Location should automatically stop the
>> destroy action if a false is returned.
>>
>> before_destroy :check_assets
>>
>> protected
>>
>> def check_assets
>>   self.assets.count == 0
>> end
>Thanks - that does work, but it makes me wonder why there isn''t a
>built in way of saying to behave like this whenever there is a
>has_many relationship.
>
>Wouldn''t that be a sensible default? Otherwise you end up with
orphans
>so when you come to say list the assets Rails blows up.
>
>Is there a simple way to propogate an error up ''Some assets still
use
>this location'' or something?
>
>-Rob
>
>--
>------------------------------------------------------
>"On two occasions, I have been asked [by members of Parliament],
>''Pray, Mr. Babbage, if you put into the machine wrong figures, will
>the right answers come out?'' I am not able to rightly apprehend the
>kind of confusion of ideas that could provoke such a question."
>-- Charles Babbage (1791-1871)
>
>"98.5% of DNA is considered to be junk DNA with no known purpose.
>Maybe it''s XML tags." -- Anon
>
>http://www.robhulme.com/
>http://robhu.livejournal.com/
>_______________________________________________
>Rails mailing list
>Rails@lists.rubyonrails.org
>http://lists.rubyonrails.org/mailman/listinfo/rails
IIRC, the default behavior is to destroy all children of a record when  
it is destroyed.

Of course the best solution here would be for you to protect the record  
at the controller level too. Here are a couple of strategies you can use.

1. a helper method for the model like ''can_destroy?'' so you
can do
something like this..

def destroy
 @item = Item.find(params[:id])
 if @item.can_destroy?
   @item.destroy
   flash[:notice] = ''record destroyed''
 else
   flash[:warning] = ''record not destroyed''
 end
end

2. having the before_destroy callback raise an exception...

def destroy
 begin
   Item.destroy(params[:id])
   flash[:notice] = ''record destroyed''
 rescue
   flash[:warning] = ''record not destroyed''
 end
end

I would probably go with the first one and write a function that  
recursively tests all the children for dependents.  Otherwise, you might  
end up with a partial destroy if the rails first tries to destroy a  
dependent object that has no children before encountering one that did.

_Kevin
www.sciwerks.com

-- 
Posted with http://DevLists.com.  Sign up and save your mailbox.

> This could be a good plugin / patch to write and submit to the core team.
> It does seem like a good default behavior.
If someone could tell me how I could sensibly bubble up an exception /
error code or something that would fit into how things are done in
rails I''ll look at making a patch for core.

-- 
------------------------------------------------------
"On two occasions, I have been asked [by members of Parliament],
''Pray, Mr. Babbage, if you put into the machine wrong figures, will
the right answers come out?'' I am not able to rightly apprehend the
kind of confusion of ideas that could provoke such a question."
-- Charles Babbage (1791-1871)

"98.5% of DNA is considered to be junk DNA with no known purpose.
Maybe it''s XML tags." -- Anon

http://www.robhulme.com/
http://robhu.livejournal.com/