From: Kevin Smith <kevins14@...>
Date: 2000-12-16T13:20:45+09:00
Subject: [ruby-talk:7390] Re: Replacing nils in arrays?


>On Fri, 15 Dec 2000  07:40:01 +0900, hal9000@hypermetrics.com wrote:
>> Question for you all.
>> 
>> Suppose you have an array that is growing
>> more or less unpredictably. You assign an
>> element that is past the current end, and
>> you get several new nils as a result.
>> 
>> My problem: I want to store zeroes, not
>> nils.
>
>quick hack:
>
>
>class Array
>  attr_accessor :default_value
>
>  alias oldidx []=
>  def []=(*arg)
>    if arg.size == 3
>      return oldidx(*arg)
>    end
>  
>    start, elem = arg
>    if start > size
>      fill @default_value, size, (start - size)
>    end
>    oldidx start, elem
>  end
>end

I would prefer to avoid storing the default value everywhere 
(in case ruby is smart enough to implement sparse arrays).
Here's my stab (I'm new to dynamic languages, so I'll leave 
it to someone else to tweak this to change the actual Array 
class behavior. Actually, I'm not sure it ever should.)

class ArrayWithDefault < Array
  attr_accessor : default
  def [](*arg)
    result = super
    if !result
      result = default
    elsif result.instance_of? Array
      result.collect! do
        | value |
        value ? value : default
      end
    end
    return result
  end
end

a1 = ArrayWithDefault.new
a1[5] = 7
a1[2] = 4
puts a1[0..9]

a1.default = 3
puts a1[0..9]

Kevin