Ruby Array#own_uniq method

Question

I have implemented my own version of Ruby's Array uniq-method (Ruby docs - Class array) as a monkey patch on Array.

Method-impl.:

require 'set'
class Array
 def own_uniq
 uniq = Set.new
 self.each { uniq.add(it) }
 uniq.to_a
 end
end

Usage:

require_relative "own_uniq.rb"
example = [1, 2, 4, 8, 16, 32, 1, 4, 16]
puts example.own_uniq.inspect # -> [1, 2, 4, 8, 16, 32]

Although it seems to work: Is my implementation generally correct?

Can it be improved? Respectively: Is there a better approach?

• radarbob: The eternal debate about "what is monkey patching?" begs the question - do you intend to add a new method or replace the existing one?
• michael.zech: It's meant as a new method. But actually the only motivation writing it, was to become more familiar with Ruby. By re-implementing it's own methods.

Answer 1

Your code can be simplified to the following:

require 'set'
class Array
 def own_uniq
 Set.new(self).to_a
 end
end

Or even:

require 'set'
class Array
 def own_uniq = Set.new(self).to_a
end

But a notable way in which your own_uniq method is not an adequate replacement for Array#uniq is that it does not permit the passing of a block to specify how to determine uniqueness as seen in the linked documentation.

An implementation that does might look like:

class Array 
 def own_uniq(&proc)
 return Set.new(self).to_a unless proc
 sort { 
 |a, b| proc.call(a) <=> proc.call(b) 
 }.each_with_object([]) { |x, set|
 set << x if set.empty? || proc.call(x) != proc.call(set.last)
 }
 end
end

If calling proc can be non-trivial, you may wish to employ a Schwartzian transform approach in implementing own_uniq.