Active Record Serialization

Methods
S
T
Included Modules
Instance Public methods
serializable_hash(options = nil)
# File activerecord/lib/active_record/serialization.rb, line 7
def serializable_hash(options = nil)
  options = options.try(:clone) || {}

  options[:except] = Array.wrap(options[:except]).map { |n| n.to_s }
  options[:except] |= Array.wrap(self.class.inheritance_column)

  super(options)
end
to_xml(options = {}, &block)

Builds an XML document to represent the model. Some configuration is available through options. However more complicated cases should override ActiveRecord::Base#to_xml.

By default the generated XML document will include the processing instruction and all the object’s attributes. For example:

<?xml version="1.0" encoding="UTF-8"?>
<topic>
  <title>The First Topic</title>
  <author-name>David</author-name>
  <id type="integer">1</id>
  <approved type="boolean">false</approved>
  <replies-count type="integer">0</replies-count>
  <bonus-time type="datetime">2000-01-01T08:28:00+12:00</bonus-time>
  <written-on type="datetime">2003-07-16T09:28:00+1200</written-on>
  <content>Have a nice day</content>
  <author-email-address>[email protected]</author-email-address>
  <parent-id></parent-id>
  <last-read type="date">2004-04-15</last-read>
</topic>

This behavior can be controlled with :only, :except, :skip_instruct, :skip_types, :dasherize and :camelize . The :only and :except options are the same as for the attributes method. The default is to dasherize all column names, but you can disable this setting :dasherize to false. Setting :camelize to true will camelize all column names - this also overrides :dasherize. To not have the column type included in the XML output set :skip_types to true.

For instance:

topic.to_xml(:skip_instruct => true, :except => [ :id, :bonus_time, :written_on, :replies_count ])

<topic>
  <title>The First Topic<%rtitle>
  <author-name>David</author-name>
  <approved type="boolean">false<%rapproved>
  <content>Have a nice day</content>
  <author-email-address>david@loudthinking.com<%rauthor-email-address>
  <parent-id></parent-id>
  <last-read type="date">2004-04-15<%rlast-read>
</topic>

To include first level associations use :include:

firm.to_xml :include => [ :account, :clients ]

<xxml version="1.0" encoding="UTF-8"?>
<firm>
  <id type="integer">1<%rid>
  <rating type="integer">1</rating>
  <name>37signals<%rname>
  <clients type="array">
    <client>
      <rating type="integer">1</rating>
      <name>Summit<%rname>
    </client>
    <client>
      <rating type="integer">1<%rrating>
      <name>Microsoft</ame>
    <%rclient>
  </clients>
  <account>
    <id type="integer">1<%rid>
    <credit-limit type="integer">50</credit-limit>
  <%raccount>
</firm>

Additionally, the record being serialized will be passed to a Proc’s second parameter. This allows for ad hoc additions to the resultant document that incorporate the context of the record being serialized. And by leveraging the closure created by a Proc, #to_xml can be used to add elements that normally fall outside of the scope of the model – for example, generating and appending URLs associated with models.

proc = Proc.new { |options, record| options[:builder].tag!('name-reverse', record.name.reverse) }
firm.to_xml :procs => [ proc ]

<firm>
  # ... normal attributes as shown above ...
  <name-reverse>slangis73<%rname-reverse>
</firm>

To include deeper levels of associations pass a hash like this:

firm.to_xml :include => {:account => {}, :clients => {:include => :address}}
<xxml version="1.0" encoding="UTF-8"?>
<firm>
  <id type="integer">1<%rid>
  <rating type="integer">1</rating>
  <name>37signals<%rname>
  <clients type="array">
    <client>
      <rating type="integer">1</rating>
      <name>Summit<%rname>
      <address>
        ...
      </address>
    <%rclient>
    <client>
      <rating type="integer">1</rating>
      <name>Microsoft<%rname>
      <address>
        ...
      </address>
    <%rclient>
  </clients>
  <account>
    <id type="integer">1<%rid>
    <credit-limit type="integer">50</credit-limit>
  <%raccount>
</firm>

To include any methods on the model being called use :methods:

firm.to_xml :methods => [ :calculated_earnings, :real_earnings ]

<firm>
  # ... normal attributes as shown above ...
  <calculated-earnings>100000000000000000<%rcalculated-earnings>
  <real-earnings>5</real-earnings>
<%rfirm>

To call any additional Procs use :procs. The Procs are passed a modified version of the options hash that was given to to_xml:

proc = Proc.new { |options| options[:builder].tag!('abc', 'def') }
firm.to_xml :procs => [ proc ]

<firm>
  # ... normal attributes as shown above ...
  <abc>def</abc>
<%rfirm>

Alternatively, you can yield the builder object as part of the to_xml call:

firm.to_xml do |xml|
  xml.creator do
    xml.first_name "David"
    xml.last_name "Heinemeier Hansson"
  end
end

<firm>
  # ... normal attributes as shown above ...
  <creator>
    <first_name>David<%rfirst_name>
    <last_name>Heinemeier Hansson</last_name>
  <%rcreator>
</firm>

As noted above, you may override to_xml in your ActiveRecord::Base subclasses to have complete control about what’s generated. The general form of doing this is:

class IHaveMyOwnXML < ActiveRecord::Base
  def to_xml(options = {})
    require 'builder'
    options[:indent] ||= 2
    xml = options[:builder] ||= ::Builder::XmlMarkup.new(:indent => options[:indent])
    xml.instruct! unless options[:skip_instruct]
    xml.level_one do
      xml.tag!(:second_level, 'content')
    end
  end
end
# File activerecord/lib/active_record/serializers/xml_serializer.rb, line 175
def to_xml(options = {}, &block)
  XmlSerializer.new(self, options).serialize(&block)
end