Question

Let’s say you have a YML file like this:

en:
  thanksgiving:
    turkey: 'Turkey'
    stuffing: 'Stuffing'
  christmas:
    ham: 'Bring ham'
  thanksgiving:
    beer: 'lots of beer'

and you want to use ruby to basically read, consolidate and rewrite that YML like this:

en:
  thanksgiving:
    turkey: 'Turkey'
    stuffing: 'Stuffing'
    beer: 'lots of beer'
  christmas:
    ham: 'Bring ham'

what is the most best / most efficient way of making this happen?

Answer 1

You need to parse the file to get the node representation:

yml = YAML.parse( open('c:\temp\foo.yml'))

The yml variable contains the whole structure. Example: Typing the following prints the actual entire file content

pp yml 

After inspecting the result I was able to write a safe serializer. Add a file called config/initializers/yaml.rb

module YAML

  def YAML.safe_load(file_name)
    YAML::safe_load_node(YAML::parse(IO.read(file_name)))
  end

  def YAML.safe_load_node(input)
    case input.kind
    when :map
      {}.tap do |h|
        input.value.each do |key, node|
          k,v = key.value, YAML::safe_load_node(node)
          if (v.is_a?(Hash) and h[k].is_a?(Hash))
            h[k] = h[k].merge(v)
          elsif (v.is_a?(Array) and h[k].is_a?(Array))
            h[k] = h[k] + v
          else
            h[k] = v
          end
        end
      end
    when :seq
      input.value.map{|node| YAML::safe_load_node(node)}
    when :scalar
      input.value
    end
  end

end 

Now in the rails console:

>> y YAML::safe_load('c:/temp/test.yml')
---
en:
  christmas:
    ham: Bring ham
  thanksgiving:
    turkey: Turkey
    stuffing: Stuffing
    beer: lots of beer