抓取网页萃取网页内容的代码
日期：2013-12-20 浏览次数：20767 次
dim sUrl
sUrl="http://travel.state.gov/visa/frvi_bulletincurrent.html"
Function streamtochar(StrStream)
set stream=CreateObject("ADODB.Stream")
stream.type=1
stream.Mode=3
stream.Open
stream.Write Strstream
stream.Position= 0
stream.Type= 2
stream.Charset="gb2312"
streamtochar= stream.ReadText
stream.Close
set stream=nothing
End Function
i = i + 1
function getContentByUrl(url)
set XmlHttp = CreateObject("MSXML2.XMLHTTP")
XmlHttp.open "GET",url,false
XmlHttp.send
getContentByUrl = streamtochar(oXmlHttp.responseBody)
set XmlHttp=nothing
end function
function getRealContent(url)
sContent = getContentByUrl(url)
getRealContent=sContent
end function
html= getContentByUrl(surl)
url_start=inStr(html," " ) url_end=inStr(html," ")
url=Mid(html,url_start,url_end-url_start)
url=replace(url,"“)
Date_start=inStr(html,"Washington, D.C. ")+57
Date_end=inStr(html," A. STATUTORY")-14
Date_T=Mid(html,Date_start,Date_end-Date_start)
sUrl="http://travel.state.gov/visa/frvi_bulletincurrent.html"
Function streamtochar(StrStream)
set stream=CreateObject("ADODB.Stream")
stream.type=1
stream.Mode=3
stream.Open
stream.Write Strstream
stream.Position= 0
stream.Type= 2
stream.Charset="gb2312"
streamtochar= stream.ReadText
stream.Close
set stream=nothing
End Function
i = i + 1
function getContentByUrl(url)
set XmlHttp = CreateObject("MSXML2.XMLHTTP")
XmlHttp.open "GET",url,false
XmlHttp.send
getContentByUrl = streamtochar(oXmlHttp.responseBody)
set XmlHttp=nothing
end function
function getRealContent(url)
sContent = getContentByUrl(url)
getRealContent=sContent
end function
html= getContentByUrl(surl)
url_start=inStr(html," " ) url_end=inStr(html," ")
url=Mid(html,url_start,url_end-url_start)
url=replace(url,"“)
Date_start=inStr(html,"Washington, D.C. ")+57
Date_end=inStr(html," A. STATUTORY")-14
Date_T=Mid(html,Date_start,Date_end-Date_start)
免责声明： 本文仅代表作者个人观点，与爱易网无关。其原创性以及文中陈述文字和内容未经本站证实，对本文以及其中全部或者部分内容、文字的真实性、完整性、及时性本站不作任何保证或承诺，请读者仅作参考，并请自行核实相关内容。
