验证当Rails的accepts

本文介绍了验证当Rails的accepts_nested_attributes_for孩子没有父母集的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！ 问题描述

我想验证的时候访问我父模型在我子模型。我发现了一些有关的HAS_ONE逆属性，但我的Rails 2.3.5不承认它，所以它必须有没有把它变成释放。我不知道这是否正是我需要的，但。

我想验证有条件基于父属性的孩子。我父模型已经建立。如果孩子没有被创建时我的update_attributes对父，那么它不具有访问父。我不知道我怎么可以访问此父。它的应的容易，像parent.build_child设置子模型的PARENT_ID，为什么不建立孩子accepts_nested_attributes_for的时候做什么呢？

例如：

类家长和LT; AR   HAS_ONE：孩子   accepts_nested_attributes_for：孩子 结束 一流的子＆lt; AR   belongs_to的：父母   validates_ presence_of：名：如果=＆GT; ：some_method   高清some_method     返回self.parent.some_condition＃=＆GT;未定义的方法`some_condition'的零：NilClass   结束 结束

我的方式是标准：

＆LT;％的form_for @parent办| F | ％＆GT;   ＆LT;％f.fields_for：孩子办| C | ％＆GT;     ＆其中;％= c.name％GT;   ＆LT;％结束％GT; ＆LT;％结束％GT;

使用的更新方法

高清更新   @parent = Parent.find（PARAMS [：ID]）   @ parent.update_attributes（PARAMS [：父]）＃=＆GT;这就是我的孩子验证发生 结束

解决方案

您不能这样做，因为内存的孩子不知道它分配给父。它只能保存后，知道了。例如，

子= parent.build_child parent.child＃=＆GT;儿童 child.parent＃=＆GT;零 ＃ 但 child.parent =父 child.parent＃=＆GT;父 parent.child＃=＆GT;儿童

所以，你可以通过手动执行反向关联那种强制此行为。例如

高清child_with_inverse_assignment =（子）   child.parent =自   self.child_without_inverse_assignment =小孩 结束 高清build_child_with_inverse_assignment（*参数）   build_child_without_inverse_assignment（*参数）   child.parent =自   儿童 结束 高清create_child_with_inverse_assignment（*参数）   create_child_without_inverse_assignment（*参数）   child.parent =自   儿童 结束 alias_method_chain：孩子=：inverse_assignment alias_method_chain：build_child，：inverse_assignment alias_method_chain：create_child，：inverse_assignment

如果你真的觉得有必要的。

P.S。它没有做，现在的原因是因为它不是太容易了。它需要被明确告知如何访问父/子在每个具体病例。身份映射一个COM prehensive办法已经解决了这个问题，但对于较新的版本有：inverse_of 解决方法。像this 之一发生在新闻组。

I'm trying to access my parent model in my child model when validating. I found something about an inverse property on the has_one, but my Rails 2.3.5 doesn't recognize it, so it must have never made it into the release. I'm not sure if it's exactly what I need though.

I want to validate the child conditionally based on parent attributes. My Parent model has already been created. If the child hasn't been created when I update_attributes on the parent, then it doesn't have access to the parent. I'm wondering how I can access this parent. It should be easy, something like parent.build_child sets the parent_id of the child model, why is it not doing it when building the child for accepts_nested_attributes_for?

For Example:

class Parent < AR has_one :child accepts_nested_attributes_for :child end class Child < AR belongs_to :parent validates_presence_of :name, :if => :some_method def some_method return self.parent.some_condition # => undefined method `some_condition' for nil:NilClass end end

My form is standard:

<% form_for @parent do |f| %> <% f.fields_for :child do |c| %> <%= c.name %> <% end %> <% end %>

With an update method

def update @parent = Parent.find(params[:id]) @parent.update_attributes(params[:parent]) # => this is where my child validations take place end

解决方案

You cannot do this because in-memory child doesn't know the parent its assigned to. It only knows after save. For example.

child = parent.build_child parent.child # => child child.parent # => nil # BUT child.parent = parent child.parent # => parent parent.child # => child

So you can kind of force this behavior by doing reverse association manually. For example

def child_with_inverse_assignment=(child) child.parent = self self.child_without_inverse_assignment = child end def build_child_with_inverse_assignment(*args) build_child_without_inverse_assignment(*args) child.parent = self child end def create_child_with_inverse_assignment(*args) create_child_without_inverse_assignment(*args) child.parent = self child end alias_method_chain :"child=", :inverse_assignment alias_method_chain :build_child, :inverse_assignment alias_method_chain :create_child, :inverse_assignment

If you really find it necessary.

P.S. The reason it's not doing it now is because it's not too easy. It needs to be explicitly told how to access parent/child in each particular case. A comprehensive approach with identity map would've solved it, but for newer version there's :inverse_of workaround. Some discussions like this one took place on newsgroups.