Hash.uniq method. Here is an example:
1 2 3 4 5 6 | [{"id"=>667824693}, {"id"=>667824693}].uniq # => [{"id"=>667824693}, {"id"=>667824693}] [{"id"=>66782469}, {"id"=>66782469}].uniq # => [{"id"=>66782469}] [{"id"=>6678246931}, {"id"=>6678246931}].uniq # => [{"id"=>6678246931}] |
Check the first command result. Very disappointing, right? So what happen? Quick looking through the Ruby code completely explained it. Here is how this method works internally (this is just prototype in Ruby, original code is in C, but works in the same way):
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | def ary_make_hash(ary, val) ary.inject({}) do |hash, el| hash[el] = val hash end end def uniq(ary) ary = ary.dup ary.uniq! ary end def uniq!(ary) hash = ary_make_hash(ary, 0) return nil if ary.length == hash.length j = 0 (0...ary.length).each do |idx| if hash.delete(ary[idx]) ary[j] = ary[idx] j += 1 end end ary.slice!(0, j) ary end |
Let’s test it:
1 2 3 4 5 6 | uniq([{"id"=>667824693}, {"id"=>667824693}]) # => [{"id"=>667824693}, {"id"=>667824693}] uniq([{"id"=>66782469}, {"id"=>66782469}]) # => [{"id"=>66782469}] uniq([{"id"=>6678246931}, {"id"=>6678246931}]) # => [{"id"=>6678246931}] |
And just to make sure our conclusions are correct:
1 2 3 4 5 6 | [{"id"=>667824693}, {"id"=>667824693}].map { |el| el.hash } # => [29793216, 29793156] [{"id"=>66782469}, {"id"=>66782469}].map { |el| el.hash } # => [255119887, 255119887] [{"id"=>6678246931}, {"id"=>6678246931}].map { |el| el.hash } # => [482552381, 482552381] |
So the idea behind the Hash.uniq method is the method Hash.hash, which produces different results for hashes in the first example. Be careful when doing obj.uniq on complex objects.
Update: There is a good research on Hash.hash method here.