執行個體講解asp抓取網上房產資訊

來源:互聯網
上載者:User

附:抓取資訊的詳細頁面案例

 

<%@LANGUAGE="VBSCRIPT" CODEPAGE="936"%>
<!-- #include file="conn.asp" -->
<!-- #include file="inc/function.asp" -->
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<title>Untitled Document</title>
<meta http-equiv="Content-Type" content="text/html; charset=gb2312">
<meta http-equiv="refresh" content="300;URL=steal_house.asp">
</head>
<body>
<%
on error resume next
'
Server.ScriptTimeout = 999999
'========================================================
'字元編碼函數
'====================================================
Function BytesToBstr(body,code)
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 =code
BytesToBstr = objstream.ReadText
objstream.Close
set objstream = nothing
End Function
'取行字串在另一字串中的出現位置
Function Newstring(wstr,strng)
Newstring=Instr(lcase(wstr),lcase(strng))
if Newstring<=0 then Newstring=Len(wstr)
End Function
'替換字串函數
function ReplaceStr(ori,str1,str2)
ReplaceStr=replace(ori,str1,str2)
end function
'====================================================
function ReadXml(url,code,start,ends)
set oSend=createobject("Microsoft.XMLHTTP")
SourceCode = oSend.open ("GET",url,false)
oSend.send()
ReadXml=BytesToBstr(oSend.responseBody,code )
start=Instr(ReadXml,start)
ReadXml=mid(ReadXml,start)
ends=Instr(ReadXml,ends)
ReadXml=left(ReadXml,ends-1)
end function
function SubStr(body,start,ends)
start=Instr(body,start)
SubStr=mid(body,start+len(start)+1)
ends=Instr(SubStr,ends)
SubStr=left(SubStr,ends-1)
end function
dim getcont,NewsContent
dim url,title
url="http://www.***.com"'新聞網址
getcont=ReadXml(url,"gb2312","<table class=k2 border=""0""","</table>")
getcont=RegexHtml(getcont)
dim KeyId,NewsClass,City,Position,HouseType,Level,Area,Price,Demostra
dim ContactMan,Contact
for i=2 to ubound(getcont)
response.Write(getcont(i)&"__<br>")

tempLink=mid(getcont(i),instr(getcont(i),"href=""")+6,instr(getcont(i),"""
onClick")-10)
tempLink=replace(tempLink,"../","")

response.Write(i&":"&tempLink&"<br>")
NewsContent=ReadXml(tempLink,"gb2312","<td valign=""bottom""
width=""400"">","<hr width=""760""
noshade size=""1"" color=""#808080"">
")
NewsContent=RemoveHtml(NewsContent)
NewsContent=replace(NewsContent,VbCrLf,"")
NewsContent=replace(NewsContent,vbNewLine,"")
NewsContent=replace(NewsContent," ","")
NewsContent=replace(NewsContent," ","")
NewsContent=replace(NewsContent," ","")
NewsContent=replace(NewsContent,"\n","")
NewsContent=replace(NewsContent,chr(10),"")
NewsContent=replace(NewsContent,chr(13),"")
'===============get Content=======================
response.Write(NewsContent)
KeyId=SubStr(NewsContent,"列號:","資訊類別:")
NewsClass=SubStr(NewsContent,"類別:","所在城市:")
City=SubStr(NewsContent,"城市:","房屋具體位置:")
Position=SubStr(NewsContent,"位置:","房屋類型:")
HouseType=SubStr(NewsContent,"類型:","樓層:")
Level=SubStr(NewsContent,"樓層:","使用面積:")
Area=SubStr(NewsContent,"面積:","樓價:")
Price=SubStr(NewsContent,"樓價:","其他說明:")
Demostra=SubStr(NewsContent,"說明:","連絡人:")
ContactMan=SubStr(NewsContent,"連絡人:","連絡方式:")
Contact=SubStr(NewsContent,"連絡方式:","資訊來源:")
response.Write("總序號:"&KeyId&"<br>")
response.Write("資訊類別:"&NewsClass&"<br>")
response.Write("所在城市:"&City&"<br>")
response.Write("房屋具體位置:"&Position&"<br>")
response.Write("房屋類型:"&HouseType&"<br>")
response.Write("樓層:"&Level&"<br>")
response.Write("使用面積:"&Area&"<br>")
response.Write("樓價:"&Price&"<br>")
response.Write("其他說明:"&Demostra&"<br>")
response.Write("連絡人:"&ContactMan&"<br>")
response.Write("連絡方式:"&Contact&"<br>")
'title=RemoveHTML(aa(i))
'response.Write("title:"&title)
for n=0 to application.Contents.count
if(application.Contents(n)=KeyId) then
ifexit=true
end if
next
if not ifexit then
application(time&i)=KeyId
'添加到資料庫
'====================================================
set rs=server.CreateObject("adodb.recordset")
rs.open "select top 1 * from news order by id desc",conn,3,3
rs.addnew
rs("NewsClass")=NewsClass
rs("City")=City
rs("Position")=Position
rs("HouseType")=HouseType
rs("Level")=Level
rs("Area")=Area
rs("Price")=Price
rs("Demostra")=Demostra
rs("ContactMan")=ContactMan
rs("Contact")=Contact
rs.update
rs.close
set rs=nothing
end if
'==================================================

next
function RemoveTag(body)
Set regEx = New RegExp
regEx.Pattern = "<[a].*?<\/[a]>"
regEx.IgnoreCase = True
regEx.Global = True
Set Matches = regEx.Execute(body)
dim i,arr(15),ifexit
i=0
j=0
For Each Match in Matches
TempStr = Match.Value
TempStr=replace(TempStr,"<td>","")
TempStr=replace(TempStr,"</td>","")
TempStr=replace(TempStr,"<tr>","")
TempStr=replace(TempStr,"</tr>","")
arr(i)=TempStr
i=i+1
if(i>=15) then
exit for
end if
Next
Set regEx=nothing
Set Matches =nothing
RemoveTag=arr

end function
function RegexHtml(body)
dim r_arr(47),r_temp
Set regEx2 = New RegExp
regEx2.Pattern ="<a.*?<\/a>"
regEx2.IgnoreCase = True
regEx2.Global = True
Set Matches2 = regEx2.Execute(body)
iii=0
For Each Match in Matches2

r_arr(iii)=Match.Value

iii=iii+1
Next
RegexHtml=r_arr
set regEx2=nothing
set Matches2=nothing
end function
'======================================================
conn.close
set conn=nothing
%>
</body>
</html>


 function.asp

<%
'**************************************************
'函數名:gotTopic
'作 用:截字串,漢字一個算兩個字元,英文算一個字元
'參 數:str ----原字串
' strlen ----截取長度
'傳回值:截取後的字串
'**************************************************
function gotTopic(str,strlen)
if str="" then
gotTopic=""
exit function
end if
dim l,t,c, i
str=replace(replace(replace(replace(str," "," "),""",chr(34)),">",">"),"<","<")
str=replace(str,"?","")
l=len(str)
t=0
for i=1 to l
c=Abs(Asc(Mid(str,i,1)))
if c>255 then
t=t+2
else
t=t+1
end if
if t>=strlen then
gotTopic=left(str,i) & "…"
exit for
else
gotTopic=str
end if
next
gotTopic=replace(replace(replace(replace(gotTopic," "," "),chr(34),"""),">",">"),"<","<")
end function
'=========================================================
'函數:RemoveHTML(strHTML)
'功能:去除HTML標記
'參數:strHTML --要去除HTML標記的字串
'=========================================================
Function RemoveHTML(strHTML)
Dim objRegExp, Match, Matches
Set objRegExp = New Regexp
objRegExp.IgnoreCase = True
objRegExp.Global = True
'取閉合的<>
objRegExp.Pattern = "<.+?>"
'進行匹配
Set Matches = objRegExp.Execute(strHTML)
' 遍曆匹配集合,並替換掉匹配的項目
For Each Match in Matches
strHtml=Replace(strHTML,Match.Value,"")
Next
RemoveHTML=strHTML
Set objRegExp = Nothing
set Matches=nothing
End Function
%>


  conn.asp
<%
'on error resume next
set conn=server.CreateObject("adodb.connection")
con= "driver={Microsoft Access Driver (*.mdb)};dbq=" & Server.MapPath("stest.mdb")

conn.open con
sub connclose
conn.close
set conn=nothing
end sub
%>



相關文章

E-Commerce Solutions

Leverage the same tools powering the Alibaba Ecosystem

Learn more >

Apsara Conference 2019

The Rise of Data Intelligence, September 25th - 27th, Hangzhou, China

Learn more >

Alibaba Cloud Free Trial

Learn and experience the power of Alibaba Cloud with a free trial worth $300-1200 USD

Learn more >

聯繫我們

該頁面正文內容均來源於網絡整理,並不代表阿里雲官方的觀點,該頁面所提到的產品和服務也與阿里云無關,如果該頁面內容對您造成了困擾,歡迎寫郵件給我們,收到郵件我們將在5個工作日內處理。

如果您發現本社區中有涉嫌抄襲的內容,歡迎發送郵件至: info-contact@alibabacloud.com 進行舉報並提供相關證據,工作人員會在 5 個工作天內聯絡您,一經查實,本站將立刻刪除涉嫌侵權內容。