R 언어 파충류 실천(상): 중고 주택 데이터 수집

주요 참고문: [1]https://www.jianshu.com/p/c092d57d275f] [2]https://blog.csdn.net/u013421629/article/details/76803903?locationNum=8&fps=1

gurl % html_session %>% 
    read_html(encoding="utf-8") %>% 
    html_nodes("div.f-main-list>div>div")
  #  puid
  puid % html_attr("puid")
  #  
  itemURL % html_attr("href") %>% gsub(pattern="/fang5",replacement="http://cs.ganji.com/fang5")
  #  
  smallImg % html_nodes("dl>dt>div>a>img") %>% html_attr("src")
  #  
  iTitle % html_nodes("dl>dd>a") %>% html_attr("title")
  #  -- 
  # ：f_mew_list > div.f-main.f-clear.f-w1190 > div.f-main-left.f-fl.f-w980 > div.f-main-list > div > div:nth-child(1) > dl > dd.dd-item.size > span:nth-child(1)
  iLayout % html_nodes("dl > dd.dd-item.size > span:nth-child(1)") %>% html_text
  #  -- 
  # ：f_mew_list > div.f-main.f-clear.f-w1190 > div.f-main-left.f-fl.f-w980 > div.f-main-list > div > div:nth-child(1) > dl > dd.dd-item.size > span:nth-child(3)
  iArea % html_nodes("dl > dd.dd-item.size > span:nth-child(3)") %>% html_text
  #  
  iOrientation% html_nodes("dl > dd.dd-item.size > span:nth-child(5)") %>% html_text#  
  iFloor % html_nodes("dl > dd.dd-item.size > span:nth-child(7)") %>% html_text#  
  #  
  iAddr % html_nodes("dl>dd>span.area") %>% html_text %>% gsub(pattern="
",replacement=" ") %>% gsub(pattern=" ",replacement="")
  #  
  iPrice % html_nodes("dl>dd>div.price>span:first-child") %>% html_text
  #  
  iunitPrice % html_nodes("dl>dd>div.time") %>% html_text %>% gsub(pattern="[^0-9]",replacement="") %>% as.numeric
  #  
  iData