This project is archived and is in readonly mode.
Accessing associations via x_ids no longer uses association conditions
Reported by Yuval Kordov | December 11th, 2010 @ 11:51 PM
I'm upgrading my Rails app from 2.3.5 to 3.0.3 and have found the following:
user.rb
has_many :relationships
has_many :contacts, :through => :relationships, :conditions
=> {:relationships => {:status => "1"}}
In Rails 2.3.5, I could call user.contact_ids and it would run through the entire condition:
user.contact_ids =>
SELECT users.id FROM users INNER JOIN
relationships ON users.id =
relationships.contact_id WHERE
((relationships.user_id = 2) AND
((relationships.status = 1)))
In Rails 3.0.3, it seems to be ignoring the parameters altogether.
user.contact_ids =>
SELECT DISTINCT relationships.contact_id FROM
relationships WHERE
(relationships.user_id = 2)
Hopefully I'm missing something obvious here, but this change has put a serious wrench in my code, where I was using _ids associations for efficiency when I only ever needed IDs.
Comments and changes to this ticket
-
Yuval Kordov December 12th, 2010 @ 10:58 PM
Forgot to add that this is using Ruby 1.9.2 and the mysql2 gem.
-
-
2kan December 19th, 2010 @ 02:59 PM
- Tag changed from “rails 3.0.3, associations, has_many” to “rails 3.0.3, associations, bug, has_many_through”
Have the same bug in edge. It works correct for has_many but incorrect for has_many :through
-
Jon Leighton December 22nd, 2010 @ 10:37 AM
- State changed from “new” to “open”
- Assigned user set to “Aaron Patterson”
- Importance changed from “” to “Low”
This problem was introduced by this optimisation: https://github.com/rails/rails/commit/373b053dc8b99dac1abc3879a17a2...
The optimisation could still work, but we'd need to be very cautious about when it's applied. For example, it could never be applied when there are
:conditionsor:orderoptions on the reflection, through_reflection, or source_reflection.My preference is to just revert the optimisation, I don't think it's worth muddying the code to tests for all these conditions to apply a small optimisation in a very specific case.
I've attached a patch with 1) a test which fails on master and 2) a commit which reverts the above, causing the test to pass again.
-
Jon Leighton December 22nd, 2010 @ 10:38 AM
- no changes were found...
-
Repository December 23rd, 2010 @ 11:21 PM
- State changed from “open” to “resolved”
(from [1619c2435b2b9c821b2b0dcab9624dbb6b23eaaa]) Revert "Optimize _ids for hm:t with belongs_to source". The optimisation has too many edge cases, such as when the reflection, source reflection, or through reflection has conditions, orders, etc. [#6153 state:resolved]
This reverts commit 373b053dc8b99dac1abc3879a17a2bf8c30302b5.
Conflicts:
activerecord/lib/active_record/associations.rbhttps://github.com/rails/rails/commit/1619c2435b2b9c821b2b0dcab9624...
Create your profile
Help contribute to this project by taking a few moments to create your personal profile. Create your profile »
<h2 style="font-size: 14px">Tickets have moved to Github</h2>
The new ticket tracker is available at <a href="https://github.com/rails/rails/issues">https://github.com/rails/rails/issues</a>
People watching this ticket
Jeremy Kemper
Jon Leighton
Michał Łomnicki
Piotr Sarnacki
Yuval Kordov
Attachments
Referenced by
-
6153
Accessing associations via x_ids no longer uses association conditions
(from [1619c2435b2b9c821b2b0dcab9624dbb6b23eaaa])
Revert ...