uniqueness of keys in group_by

[cribe]

Ok, I have a class that extends array but I'm seeing bizarre behavior when I try to do group_by on it.

# first I thought, maybe it uses object_id
"a".object_id == "a".object_id #=> false
["a", "b", "a"].group_by {|i| i} #=> #<OrderedHash {"a"=>["a", "a"], "b"=>["b"]}>

#ok, that's not it. maybe inspect()?
class Klass
def initialize(v)
@value = v
end
def inspect
@value
end
end
arr = []
(0..10).each do |x|
arr << Klass.new(rand(3))
end
arr.group_by {|i| i} #=> #<OrderedHash {2=>[2], 1=>[1], 0=>[0], 2=>[2], 1=>[1], 1=>[1], 2=>[2], 0=>[0], 2=>[2], 2=>[2], 0=>[0]}>

#nope. and yet:
[2, 1, 0, 2, 1, 1, 2, 0, 2, 2, 0].group_by {|i| i} #=>#<OrderedHash {0=>[0, 0, 0], 1=>[1, 1, 1], 2=>[2, 2, 2, 2, 2]}>

# however, if I explicitly call inspect then it works
arr.group_by {|i| i.inspect} #=> #<OrderedHash {0=>[0, 0, 0], 1=>[1, 1, 1], 2=>[2, 2, 2, 2, 2]}>

Any ideas?

(BTW this was all being done in a rails console which is why the results were returning orderedhashes instead of regular hashes but the results are the same in irb anyway)

 

[Robert Klemme]

Ok, I have a class that extends array but I'm seeing bizarre behavior
when I try to do group_by on it.

# first I thought, maybe it uses object_id "a".object_id ==
"a".object_id #=> false ["a", "b", "a"].group_by {|i| i} #=>
#<OrderedHash {"a"=>["a", "a"], "b"=>["b"]}>

#ok, that's not it. maybe inspect()? class Klass def initialize(v)
@value = v end def inspect @value end end arr = [] (0..10).each do
|x| arr<< Klass.new(rand(3)) end arr.group_by {|i| i} #=>
#<OrderedHash {2=>[2], 1=>[1], 0=>[0], 2=>[2], 1=>[1], 1=>[1],
2=>[2], 0=>[0], 2=>[2], 2=>[2], 0=>[0]}>

#nope. and yet: [2, 1, 0, 2, 1, 1, 2, 0, 2, 2, 0].group_by {|i| i}
#=>#<OrderedHash {0=>[0, 0, 0], 1=>[1, 1, 1], 2=>[2, 2, 2, 2, 2]}>

# however, if I explicitly call inspect then it works arr.group_by
{|i| i.inspect} #=> #<OrderedHash {0=>[0, 0, 0], 1=>[1, 1, 1],
2=>[2, 2, 2, 2, 2]}>

Any ideas?

$ ri19 Enumerable#group_by
Enumerable#group_by

(from ruby core)
------------------------------------------------------------------------------
enum.group_by {| obj | block } -> a_hash
enum.group_by -> an_enumerator

------------------------------------------------------------------------------

Returns a hash, which keys are evaluated result from the block, and
values are
arrays of elements in enum corresponding to the key.

If no block is given, an enumerator is returned instead.

(1..6).group_by {|i| i%3} #=> {0=>[3, 6], 1=>[1, 4], 2=>[2, 5]}


Hash keys are evaluated based on #hash and #eql? methods.

Kind regards

robert