Ruby script to verify the speed of xml parsing by Nokogiri.

test_nokogiri_xml.rb

#! /usr/local/bin/ruby
# coding: utf-8
#
# Ruby script to verify the speed of xml parsing by Nokogiri.
#
# Copyright(C) 2015 mk-mode.com All Rights Reserved.
#---------------------------------------------------------------------------------
#++
require 'benchmark'
require 'nokogiri'

class TestNokogiriXml
  def initialize
    @xml  = Nokogiri::XML(File.read("atom/2015/08/27/20150827172859545_atom.xml"))
    @secs = 10000
    @ns   = {"a" => "http://www.w3.org/2005/Atom"}
  end

  def exec
    # 1. css メソッド(CSS セレクタ)による解析
    printf("CASE-1: %.4f secs.\n", case_1)

    # 2. search メソッド(CSS セレクタ)による解析
    printf("CASE-2: %.4f secs.\n", case_2)

    # 3. search メソッド(XPath, Namespace 指定なし)による解析
    printf("CASE-3: %.4f secs.\n", case_3)

    # 4. / メソッド(CSS セレクタ)による解析
    printf("CASE-4: %.4f secs.\n", case_4)

    # 5. / メソッド(XPath, Namespace 指定なし)による解析
    printf("CASE-5: %.4f secs.\n", case_5)

    # 6. xpath メソッド(XPath, Namespace 指定なし)による解析
    printf("CASE-6: %.4f secs.\n", case_6)

    # 7. xpath メソッド(XPath, Namespace 指定あり)による解析
    printf("CASE-7: %.4f secs.\n", case_7)
  rescue => e
    $stderr.puts "[#{self.class.name}.#{__method__}] #{e}"
    e.backtrace.each{ |tr| $stderr.puts "\t#{tr}" }
    exit 1
  end

  private

  def case_1
    return Benchmark.realtime do
      @secs.times do |i|
        @xml.css("entry").each do |e|
          title = e.css("title").first.text
        end
      end
    end
  rescue => e
    raise
  end

  def case_2
    return Benchmark.realtime do
      @secs.times do |i|
        @xml.search("entry").each do |e|
          title = e.search("title").first.text
        end
      end
    end
  rescue => e
    raise
  end

  def case_3
    return Benchmark.realtime do
      @secs.times do |i|
        @xml.search("//xmlns:entry").each do |e|
          title = e.search("./xmlns:title").first.text
        end
      end
    end
  rescue => e
    raise
  end

  def case_4
    return Benchmark.realtime do
      @secs.times do |i|
        (@xml/"entry").each do |e|
          title = (e/"title").first.text
        end
        # 以下でも同様。しかし、ごく少し遅い。
        #(@xml/:entry).each do |e|
        #  title = (e/:title).first.text
        #end
      end
    end
  rescue => e
    raise
  end

  def case_5
    return Benchmark.realtime do
      @secs.times do |i|
        (@xml/"//xmlns:entry").each do |e|
          title = (e/"./xmlns:title").first.text
        end
      end
    end
  rescue => e
    raise
  end

  def case_6
    return Benchmark.realtime do
      @secs.times do |i|
        @xml.xpath("//xmlns:entry").each do |e|
          title = e.xpath("xmlns:title").first.text
        end
      end
    end
  rescue => e
    raise
  end

  def case_7
    return Benchmark.realtime do
      @secs.times do |i|
        @xml.xpath("//a:entry", @ns).each do |e|
          title = e.xpath("a:title", @ns).first.text
        end
      end
    end
  rescue => e
    raise
  end
end

TestNokogiriXml.new.exec