hpricot vs nokogiry

yakczh

　同一个url ,同一个xpath

url='http://zu.cq.soufun.com/house/c21000-d22000-g22-s31-kw%bd%f0%c9%bd%c3%fb%b6%bc/'
xpath="//p[@class='housetitle']/a"
   

用hpricot 抓取

1. require 'hpricot'
   
2. 
   
3. doc = Hpricot(open(url))
   
4. 
   
5. doc.search(xpath).each do |item|
   
6.   puts item['href']
   
7. end

复制代码

有结果　

用nokogiri 抓取

1. require 'nokogiri'
   
2. doc = Nokogiri::HTML(open(url))
   
3.        
   
4. #    puts doc
   
5. doc.xpath(xpath).each do |link|
   
6.     puts link.content
   
7.    puts link['href']
   
8.          end

复制代码

无结果　

zhlong8

模块的名字很奇葩

yakczh

回复 2# Sevk

1. 
   
2. require 'pp'
   
3. require 'open-uri'
   
4. require 'nokogiri'
   
5. 
   
6. url='http://zu.cq.soufun.com/house/c21000-d22000-g22-s31-kw%bd%f0%c9%bd%c3%fb%b6%bc/'
   
7. xpath="//p[@class='housetitle']/a"
   
8.    
   
9. 
   
10. doc = Nokogiri::HTML(open(url))
    
11.        
    
12. #    puts doc
    
13. doc.xpath(xpath).each do |link|
    
14.     puts link.content
    
15.    puts link['href']
    
16.    pp link
    
17.          end

复制代码

还是没有

yakczh

1. doc.css("p.housetitle").each do |link|
   
2.         puts link
   
3.         
   
4.         end

复制代码

用css选择器也取不出来

yakczh

用hpricot抓取　如果要输出　链接的内部文本　比如

  text= link.inner_text
puts text

报错　Ruby192/lib/ruby/gems/1.9.1/gems/hpricot-0.8.6-x86-mswin32/lib/hpricot/builder.rb:9:in `gsub': invalid byte sequence in UTF-8 (ArgumentError)

要强制转码才行

nokogiri可以直接在构造函数中传编码参数　　  doc = Nokogiri::HTML(open(url),'gbk')  
hpricot好象没用，只认utf8