just curious about CSS and XPath selector performance. did some benchmarking

css_xpath_selector_benchmarking.rb

require 'rubygems'
require 'mechanize'
require 'benchmark'

agent = Mechanize.new
# however, this fetch already spend around 2 seconds. one http request response.
page = agent.get('http://www.ufood.com.hk/search/search.action?name=&distIds=4,10,11,3,5,8,6,7,2,1,9,12,13,14,15,18,16,17,19,20,21,22,23,24,25,26,33,34,27,35,32,30,29,28,31&ftIds=')

Benchmark.bm do |b| 
  # search by tag name and class 10.470000   0.040000  10.510000 ( 11.149709)
  b.report("search by tag name and class") {
    5_000.times {
      page.search('div.footer')
    }
  }
  # search by xpath  0.890000   0.030000   0.920000 (  0.967669)
  b.report("search by xpath") {
    5_000.times {
      page.search('//center/div[4]')
    }
  }
  
  # search by full xpath  0.330000   0.020000   0.350000 (  0.359641)
  b.report("search by full xpath") {
    5_000.times {
      page.search('/html/body/center/div[4]')
    }
  }
  
  # search by class 39.860000   0.150000  40.010000 ( 44.043039)
  b.report("search by class") {
    5_000.times {
      page.search('.footer')
    }
  }  
end