Question

ruby 1.8.7 (2008-08-11 patchlevel 72) [i586-linux]

From script/console:

>> pairs = Array.new(2).map!{Array.new(2).map!{Array.new(2, Array.new)}}
=> [[[[], []], [[], []]], [[[], []], [[], []]]]
>> pair = Pair.first
=> #<Pair id: 39, charge_card_id: 1, classroom_id: 1, timeslot_id: 1, created_at: "2010-04-01 00:45:37", updated_at: "2010-04-01 00:45:47">
>> pairs[0][0][0] << pair
=> [#<Pair id: 39, charge_card_id: 1, classroom_id: 1, timeslot_id: 1, created_at: "2010-04-01 00:45:37", updated_at: "2010-04-01 00:45:47">]
>> pairs[0][0]
=> [[#<Pair id: 39, charge_card_id: 1, classroom_id: 1, timeslot_id: 1, created_at: "2010-04-01 00:45:37", updated_at: "2010-04-01 00:45:47">], [#<Pair id: 39, charge_card_id: 1, classroom_id: 1, timeslot_id: 1, created_at: "2010-04-01 00:45:37", updated_at: "2010-04-01 00:45:47">]]
>>

So the question is why the pair object appears in pairs[0][0][0] AND in pairs[0][0][1] inspite of I did’n ask it to appear there. Notice I don’t nedd to pairs[0][0][0] = pair – I want it as a first array member, so I need to use << or .push.

Answer 1

First of all, you want

pairs = Array.new(2) { Array.new(2) { Array.new(2) { [] }}}

instead of what you got. Two major differences:

• you save yourself the #map! calls
• in your example, “Array.new(2, Array.new)” is creating one Array which is used for both indices, so you are refering to the same array twice. By using the block syntax, you are ensuring that for every index you are having one separate instance of Array