使用瀏覽器打開要采集的頁面(如:http://sports.sina.com.cn/k/2008-09-15/04593948756.shtml,你可以其他頁面),打開后,點擊右鍵,查源文件。
< %
'功能:asp采集代碼
'作者:wangsdong
'備注:支持原創程序,請保留此信息,謝謝
url="http://sports.sina.com.cn/k/2008-09-15/04593948756.shtml"
str=getHTTPPage(url)
title=strcut(str,"<h1 id=""artibodyTitle"" style=""color:#03005C;"">","</h1>",2)
content=strcut(str,"<!-- 正文內容 begin -->","<!-- 正文內容 end -->",2)
response.write "新聞標題<br><b>"&title&"</b><br><br><br>新聞內容:<br>"&content
Function getHTTPPage(url)
On Error Resume Next
dim http
set http=Server.createobject("Microsoft.XMLHTTP")
Http.open "GET",url,false
Http.send()
if Http.readystate<>4 then
exit function
end if
getHTTPPage=bytesToBSTR(Http.responseBody,"GB2312")
set http=nothing
If Err.number<>0 then
Response.Write "<p align='center'><font color='red'><b>
服務器獲取文件內容出錯</b></font></p>"
Err.Clear
End If
End Function
Function BytesToBstr(body,Cset)
dim objstream
set objstream = Server.CreateObject("adodb.stream")
objstream.Type = 1
objstream.Mode =3
objstream.Open
objstream.Write body
objstream.Position = 0
objstream.Type = 2
objstream.Charset = Cset
BytesToBstr = objstream.ReadText
objstream.Close
set objstream = nothing
End Function
'截取字符串,1.包括起始和終止字符,2.不包括
Function strCut(strContent,StartStr,EndStr,CutType)
Dim strHtml,S1,S2
strHtml = strContent
On Error Resume Next
Select Case CutType
Case 1
S1 = InStr(strHtml,StartStr)
S2 = InStr(S1,strHtml,EndStr)+Len(EndStr)
Case 2
S1 = InStr(strHtml,StartStr)+Len(StartStr)
S2 = InStr(S1,strHtml,EndStr)
End Select
If Err Then
strCute = "<p align='center'>沒有找到需要的內容。</p>"
Err.Clear
Exit Function
Else
strCut = Mid(strHtml,S1,S2-S1)
End If
End Function
% >