Optimizing creating new classes into api

In one of my Rails applications I'm working with a huge amount of data from Amazon AWS S3. Currently this is my code where I am creating a new Batch and looping through a Hash created in my Batch class.

Class User < ActiveRecord::Base
...
  def create_batch
    batch = Batch.new
    obj = batch.access_bucket
    i = 0

    while i < obj[:folder].length do
      @batch = Batch.new
      @batch.folder = obj[:folder][i]
      @batch.batch_number = obj[:batch_number][i]
      @batch.url = obj[:url][i]
      @batch.save!
      i = i+1
    end
  end
end

The Batch class looks as such:

class Batch < ActiveRecord::Base
  before_create :access_bucket
  Obj = {:folder => [], :batch_number => [], :url => []}
  def access_bucket
    s3 = AWS::S3.new
    bucket = s3.buckets['curateanalytics']
    bucket.objects.each do |obj|
      if obj.key =~ /swipe batches/
        if obj.key =~ /jpg/
          Obj[:folder] << sort_objs(obj.key)[0]
          Obj[:batch_number] << sort_objs(obj.key)[1]
          Obj[:url] << sort_objs(obj.key)[2]
        end
      end
    end
    return Obj
  end

  def sort_objs(url)
    swipe = url.split("swipe batches/").last.split("/").first
    batch_id = url.split("swipe batches/").last.split("/")[1]
    file = url.split("swipe batches/").last.split("/").last
    return [swipe, batch_id, file]
  end
end

The problem is each key of the Obj Hash contains 3402 elements, most of which are the same. While my initial sorting is parsing so I can see what the JSON will look like my two questions are

1) is there anyway to optimize this code so it will not need to loop through all 3402 elements?

2) Can I use .uniq on the @batch.folder so that each batch with the same name will be filed under that JSON subcategory?