ファイルの文字コード変換のメモリ使用量を調べる、的な

gistfile1.rb

require 'objspace'
require 'csv'
require 'nkf'
require 'tempfile'

def test_a
  src_encoding = NKF.guess(File.read(ARGV[0], 1_048_576))
  case src_encoding
  when Encoding::US_ASCII, Encoding::UTF_8
    'UTF-8'
  when Encoding::Shift_JIS, Encoding::WINDOWS_31J
    'CP932'
  when Encoding::EUC_JP
    'EUC-JP'
  else
    raise "Invalid encoding of source_file."
  end

  open(ARGV[0], "r:#{src_encoding}:UTF-8", undef: :replace, replace: '') do |f|
    CSV.new(f, col_sep: '|').each do |row|
      puts `ps -o rss= -p #{Process.pid}`.to_i
    end
  end
end

def test_b
  Tempfile.open("encoded_file") do |temp|
    File.open(ARGV[0]) do |f|
      while !f.eof?
        chunk = f.read(1_048_576)

        while (one_byte = f.read(1)) != "\n" && !f.eof?
          chunk << one_byte
        end

        temp.puts chunk.encode('UTF-8', 'CP932', undef: :replace, replace: '')
        chunk.encode!('UTF-8', 'CP932', undef: :replace, replace: '')
        $stdout.puts `ps -o rss= -p #{Process.pid}`.to_i
        chunk.clear
      end
    end
  end
end

# puts ObjectSpace.memsize_of_all
puts `ps -o rss= -p #{Process.pid}`.to_i

test_b

# puts ObjectSpace.memsize_of_all
puts `ps -o rss= -p #{Process.pid}`.to_i