Rails - Aug 2006 - HABTM and getting to arrays from either direction

I didn''t find exactly an answer on this one, so I''d like to
get a
clarification from folks who live & breathe HABTM.

class foo
  has_and_belongs_to_many :bars
end

class bar
  has_and_belongs_to_many :foos
end

(and yes, I have a bars_foos table in my db)

def test_habtm
  bar = Bar.new
  foo = Foo.new

  foo.bars<<bar
  assert !foo.bars.empty?
  assert !bar.foos.empty?
end

My first assert passes just fine.  But the 2nd fails.  Am I wrong on the 
concepts of HABTM? Shouldn''t the << operator add to the join
table and
therefore the next find from either object return a non-empty array?

thx,
--dwf



-- 
Posted via http://www.ruby-forum.com/.

DWFrank wrote:
> I didn''t find exactly an answer on this one, so I''d like
to get a
> clarification from folks who live & breathe HABTM.
> 
> class foo
>   has_and_belongs_to_many :bars
> end
> 
> class bar
>   has_and_belongs_to_many :foos
> end
> 
> (and yes, I have a bars_foos table in my db)
> 
> def test_habtm
>   bar = Bar.new
>   foo = Foo.new
> 
>   foo.bars<<bar
>   assert !foo.bars.empty?
>   assert !bar.foos.empty?
> end
> 
> My first assert passes just fine.  But the 2nd fails.  Am I wrong on the 
> concepts of HABTM? Shouldn''t the << operator add to the join
table and
> therefore the next find from either object return a non-empty array?
<< acts different depending on whether the record has been saved or not. 
If it is not a new record (has already been saved), the join table will 
be updated when << is sent. If it hasn''t yet been saved, it has
no id to
put in the join table, so it just saves the associated object in the 
bars collection and waits to create the entry in the join table when it 
gets saved. Try saving the objects and see what happens then.

--
Josh Susser
http://blog.hasmanythrough.com/

-- 
Posted via http://www.ruby-forum.com/.

Figured out my problem. Validation was preventing the save of a foo 
within the test code.

I did the smart thing and made EXACTLY the code above and it 
bar.foos.empty? is true.

What I did find odd was that bar.foos.include?(foo) doesn''t work. 
I''m
sure this has something to do with ActiveRecord arrays.  But I''m now 
using bar.foos.find(foo.id) and I''m happy.

thx,
--dwf

-- 
Posted via http://www.ruby-forum.com/.

DWFrank wrote:
> I didn''t find exactly an answer on this one, so I''d like
to get a
> clarification from folks who live & breathe HABTM.
> 
> class foo
>   has_and_belongs_to_many :bars
> end
> 
> class bar
>   has_and_belongs_to_many :foos
> end
> 
> (and yes, I have a bars_foos table in my db)
> 
> def test_habtm
>   bar = Bar.new
>   foo = Foo.new
> 
>   foo.bars<<bar
>   assert !foo.bars.empty?
>   assert !bar.foos.empty?
> end
> 
> My first assert passes just fine.  But the 2nd fails.  Am I wrong on the 
> concepts of HABTM? Shouldn''t the << operator add to the join
table and
> therefore the next find from either object return a non-empty array?
Not an HABTM expert but...

The foos.bars collection is not empty because it''s an in-memory array
to which you''ve explicitly appended bar.  The bar.foos collection
won''t include foo until foo is saved (also saving bar and creating
the join table entry).

I suppose Rails could be written to automatically keep the reverse
collection in sync, both prior to and through saves, but the code
would be a bit tricky.

-- 
We develop, watch us RoR, in numbers too big to ignore.