rails: ActiveModel#Errors creates key in messages hash if there are any errors for that given key

If you want to access an error for an attribute with model.errors[:foo] if there are any error on that attribute, it creates an entry on messages hash with this attribute name. So after try to access errors of an attribute if you want to get all error keys from the ActiveModel::Errors instance it will return wrong keys. Lets look at the example;

  post = Post.new
  post.errors.keys # => []
  post.errors[:foo] # => []
  post.errors.keys # => [:foo]

I think this is intended behaviour but it’s weird and error prone. Is there any reason to create a key on messages hash for given attribute name if there are any errors on that attribute?

About this issue

Original URL
State: closed
Created 8 years ago
Comments: 17 (13 by maintainers)

Commits related to this issue

Fix ActiveModel::Errors #keys, #values Before: person.errors.keys # => [] person.errors.values # => [] person.errors[:name] # => [] person.errors.keys # => [:name] person.errors.val... — committed to bogdanvlviv/rails by bogdanvlviv 7 years ago

Most upvoted comments

This means one cannot reliably call errors.keys or errors.messages.keys and be sure that all the contained keys are actual errors.

Exactly. But this behaviour is documented here. I know that it’s utterly weird to have side effect for a reader method but at least we have documentation about it.

One solution might be using a wrapper class around array or a class which is extended from array to store error messages which acts like falsy if it is empty.

meinac on Mar 24, 2017