# Has\_one not working as expected? **URL:** \ **Category:** Rails\ **Created:** [January 22, 2006, 4:43am UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903 "2006-01-22T04:43:21Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 22, 2006, 4:43am UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/1 "2006-01-22T04:43:21Z")
I messed around with Rails for a while last year, but then got busy and had to drop it. However, I now have time again, so I’m starting to play around with a simple database schema to model my extensive collection of photographic material, including equipment, negatives, photographs, etc. I’m running into an issue with using belongs\_to/has\_one methods. Read on for an explanation… I have two postgres database tables, defined as follows: photo\_dev=\> \d manufacturers Table “public.manufacturers” Column | Type | Modifiers --------±-----------------------±-------------------------------------------------------------- id | integer | not null default nextval(‘public.manufacturers\_id\_seq’::text) name | character varying(255) | not null Indexes: “manufacturers\_pkey” primary key, btree (id) photo\_dev=\> \d filters Table “public.filters” Column | Type | Modifiers -----------------±-----------------------±-------------------------------------------------------- id | integer | not null default nextval(‘public.filters\_id\_seq’::text) name | character varying(255) | not null filter\_factor | numeric(2,1) | not null manufacturer\_id | integer | not null description | character varying(50) | not null Indexes: “filters\_pkey” primary key, btree (id) Foreign-key constraints: “$1” FOREIGN KEY (manufacturer\_id) REFERENCES manufacturers(id) ON DELETE CASCADE The tables contain this test data: photo\_dev=\> select \* from filters; id | name | filter\_factor | manufacturer\_id | description ----±-----------±--------------±----------------±------------ 24 | Lee Filter | 3.0 | 5 | Deep Red (2 rows) photo\_dev=\> select \* from manufacturers; id | name ----±----- 5 | Lee (1 row) I’ve cobbled together this test file: #!/usr/bin/ruby require ‘pp’ require ‘rubygems’ require\_gem ‘activerecord’ ActiveRecord::Base::establish\_connection( :adapter =\> ‘postgresql’, :host =\> ‘localhost’, :database =\> ‘photo\_dev’, :username =\> ‘photo’, :password =\> ‘photo’ ) class Manufacturer \< ActiveRecord::Base belongs\_to :filter end class Filter \< ActiveRecord::Base has\_one :manufacturer, :foreign\_key =\> ‘id’ end filter = Filter.find( 24 ); pp filter print filter.manufacturer, “\n” man = Manufacturer.find( filter.manufacturer\_id ); pp man The first problem I had was that placing this statement: ``` has_one :manufacturer ``` in the Filter class could cause invalid SQL to be generated when trying to lookup the manufacturer. For example, I would have seen this SQL in the stack trace: ``` SELECT * FROM manufacturers WHERE (manufacturers.filter_id = 24) ``` LIMIT 1 the wrong column name (manufacturers.filter\_id) was used - I had to add the :foreign\_key =\> ‘id’ to fix the column name problem. Is this normal, or have I defined the tables incorrectly? It appears that the foreign key name assumed to also be the name of the primary key in the child table, which breaks the assumption that all tables have a primary key called ‘id’… If I run the above script, I get the following output: #\“Lee Filter”, “manufacturer\_id”=\>“5”, “filter\_factor”=\>“3.0”, “id”=\>“24”, “description”=\>“Deep Red”}\> nil #\“Lee”, “id”=\>“5”}\> Calling filter.manufacturer is not finding the child record, yet if I directly fetch the record, I can find it. What gives? My guess is that the two problems are related, and something is borked either in my code, or in my database definitions. Anyone have any suggestions? I’m running on Rails 1.0… Cheers! ``` -klm. ``` ---
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 22, 2006, 5:25am UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/2 "2006-01-22T05:25:34Z")
I’m using the supplied scaffold.css file and in my layout I have a transitional doctype: Using this doctype and scaffold.css, when error\_messages\_for() is used, the styles don’t get applied. I believe it is due to the css using #ErrorExplanaion and the actual id/class names are errorExplanation. Note the capitalization of the ‘E’ in error. Fixing the css allowed the styles to be applied to the default output of error\_messages\_for. Should this be fixed? Thanks. Bob S. ---
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 22, 2006, 8:14am UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/3 "2006-01-22T08:14:51Z")
Wow! My scaffolding derived forms look a _lot_ better now. Thanks! I think you need to file a bug report for this… – – Tom M. ---
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 22, 2006, 8:23am UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/4 "2006-01-22T08:23:59Z")
Bug report submitted. ---
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 22, 2006, 6:56am UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/5 "2006-01-22T06:56:28Z")
On Sat, Jan 21, 2006 at 08:39:59PM -0700, Ken Miller wrote: > Calling filter.manufacturer is not finding the child record, yet if I > directly fetch the record, I can find it. What gives? Your associations are reversed. Instead of: ``` Filter has_one :manufacturer, ... Manufacturer belongs_to :filter ``` Use: ``` Filter belongs_to :manufacturer Manufacturer has_one :filter (or, has_many :filters) ``` The belongs\_to association (whether in a one-to-one or one-to-many relationship) always applies to the class whose table has the foreign key referencing the other. In your case, your filters table has a FK referencing the manufacturer. Ergo, Filter belongs\_to Manufacturer. See the API docs[1] for more examples. Erik. [1] [http://ar.rubyonrails.org/classes/ActiveRecord/Associations/ClassMethods.html](http://ar.rubyonrails.org/classes/ActiveRecord/Associations/ClassMethods.html) ---
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 27, 2006, 1:43pm UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/6 "2006-01-27T13:43:40Z")
baz Scott wrote: > Thank you Bob, I spent hours trying to figure this out today and I got > it down to commenting out the doctype made errors work - but that just > confused me more, simple capitalisation, why is it always the simple > things that hurt the most? Because that’s when you kick yourself the hardest 🙂 ---
**Author:** ![Ken\_Miller](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/k/b2d939/32.png) [@Ken\_Miller](https://www.ruby-forum.com/u/Ken_Miller)\ **Post date:** [January 27, 2006, 1:37pm UTC](https://www.ruby-forum.com/t/has-one-not-working-as-expected/63903/7 "2006-01-27T13:37:34Z")
Thank you Bob, I spent hours trying to figure this out today and I got it down to commenting out the doctype made errors work - but that just confused me more, simple capitalisation, why is it always the simple things that hurt the most? Bob S. wrote: > I’m using the supplied scaffold.css file and in my layout I have a > transitional doctype: > > Using this doctype and scaffold.css, when error\_messages\_for() is used, > the > styles don’t get applied. I believe it is due to the css using > #ErrorExplanaion and the actual id/class names are errorExplanation. > Note > the capitalization of the ‘E’ in error. Fixing the css allowed the > styles to > be applied to the default output of error\_messages\_for. > > Should this be fixed? > > Thanks. > > Bob S.