form1.frm

提取在HTM文件中的URL地址！包括可以去重复地址功能

VERSION 5.00
Object = "{F9043C88-F6F2-101A-A3C9-08002B2F49FB}#1.2#0"; "COMDLG32.OCX"
Begin VB.Form Form1 
   Caption         =   "url连接提取----混子20070117"
   ClientHeight    =   4785
   ClientLeft      =   60
   ClientTop       =   345
   ClientWidth     =   4140
   LinkTopic       =   "Form1"
   ScaleHeight     =   4785
   ScaleWidth      =   4140
   StartUpPosition =   3  '窗口缺省
   Begin MSComDlg.CommonDialog CommonDialog1 
      Left            =   600
      Top             =   4920
      _ExtentX        =   847
      _ExtentY        =   847
      _Version        =   393216
   End
   Begin VB.CommandButton Command3 
      Caption         =   "保存"
      Height          =   495
      Left            =   2640
      TabIndex        =   5
      Top             =   3240
      Width           =   1215
   End
   Begin VB.CommandButton Command2 
      Caption         =   "去重复"
      Height          =   495
      Left            =   1440
      TabIndex        =   4
      Top             =   3240
      Width           =   1095
   End
   Begin VB.CommandButton Command1 
      Caption         =   "提取"
      Height          =   495
      Left            =   120
      TabIndex        =   3
      Top             =   3240
      Width           =   1215
   End
   Begin VB.ListBox lstemail 
      Height          =   2400
      Left            =   120
      TabIndex        =   2
      Top             =   720
      Width           =   3855
   End
   Begin VB.TextBox Text1 
      Height          =   375
      Left            =   1920
      TabIndex        =   1
      Text            =   "C:\Documents and Settings\Administrator\桌面\url"
      Top             =   120
      Width           =   2055
   End
   Begin VB.Label lblEmail 
      Height          =   375
      Left            =   120
      TabIndex        =   7
      Top             =   3840
      Width           =   3855
   End
   Begin VB.Label Label2 
      Height          =   375
      Left            =   120
      TabIndex        =   6
      Top             =   4320
      Width           =   3855
   End
   Begin VB.Label Label1 
      Caption         =   "输入HTM文件所在文件夹，注意后的\"
      Height          =   375
      Left            =   120
      TabIndex        =   0
      Top             =   120
      Width           =   1575
   End
End
Attribute VB_Name = "Form1"
Attribute VB_GlobalNameSpace = False
Attribute VB_Creatable = False
Attribute VB_PredeclaredId = True
Attribute VB_Exposed = False
Option Explicit

Dim x, Y, St1, St2, tmpY As Integer

Private Sub StripEmail(FilePath As String)
Dim tmpEmail1, tmpEmail2 As String
Open FilePath For Input As #1
Do Until EOF(1)
On Error Resume Next
Input #1, tmpEmail1  'tmpemail1为打开的HTML
For x = 1 To Len(tmpEmail1)  '循环总长度
tmpEmail2 = Mid(tmpEmail1, x, 7)  '从tmepmil第x位置取7个字符
If tmpEmail2 = "http://" Then
St1 = x  '当前循环数
tmpY = x + 1
For Y = 1 To Len(tmpEmail1)
tmpEmail2 = Mid(tmpEmail1, tmpY, 1)
If tmpEmail2 = Chr(34) Or tmpEmail2 = "/" Then
St2 = tmpY
tmpEmail2 = Mid(tmpEmail1, St1 + 7, ((St2 - St1) - 7))
If (Left(tmpEmail2, 2) <> "//") And (Left(tmpEmail2, 1) <> "/") Then
lstemail.AddItem "http://" + tmpEmail2
Exit For
End If
End If
tmpY = tmpY + 1
Next Y
End If
Next x
Loop
Close #1
End Sub
Private Sub Command1_Click()
Dim fs As New FileSystemObject ' 建立 FileSystemObject
Dim fd As Folder ' 定义 Folder 对象
Dim sfd As Folder
Set fd = fs.GetFolder(Text1)
Command1.Enabled = False
Screen.MousePointer = vbHourglass
FindFile fd, "*.htm" 'Text1.Text
Command1.Enabled = True
Screen.MousePointer = vbDefault
End Sub
Sub FindFile(fd As Folder, FileName As String)
Dim sfd As Folder, f As File
' Part I