C# 將Word，Excel轉換成Html

轉：http://skillerguo.blog.163.com/blog/static/1709513920071853219948/

 

首先，需要瞭解的是VBA編程，它是Office提供的外部編程介面，我採用的是Office2003所以添加的引用是Microsoft Office 11.0 Object Library 這個Com組件，否則無法操作Office的。下面開始：

Word -〉Html

Microsoft.Office.Interop.Word.ApplicationClass appclass = new Microsoft.Office.Interop.Word.ApplicationClass();//執行個體化一個Word
Type wordtype = appclass.GetType();
Microsoft.Office.Interop.Word.Documents docs = appclass.Documents;//擷取Document
Type docstype = docs.GetType();
object filename = Application.StartupPath+@"\ccyt_chm\inform\"+n.FullPath;//n.FullPath為Word檔案的路徑
Microsoft.Office.Interop.Word.Document doc = (Microsoft.Office.Interop.Word.Document)docstype.InvokeMember("Open",System.Reflection.BindingFlags.InvokeMethod,null,docs,new object[]{filename,true,true});//開啟檔案
Type doctype = doc.GetType();
object savefilename = Application.StartupPath+@"\ccyt_chm\inform\"+n.FullPath.Split('.').GetValue(0)+".html";
doctype.InvokeMember("SaveAs",System.Reflection.BindingFlags.InvokeMethod,null,doc,new object[]{savefilename,Microsoft.Office.Interop.Word.WdSaveFormat.wdFormatHTML});//另存新檔Html格式
wordtype.InvokeMember("Quit",System.Reflection.BindingFlags.InvokeMethod,null,appclass,null);//退出
Thread.Sleep(3000);//為了使退出完全，這裡阻塞3秒
StreamReader objreader = new StreamReader(savefilename.ToString(),System.Text.Encoding.GetEncoding("GB2312"));  //以下內容是為了在Html中加入對本身Word檔案的下載       
FileStream fs = new FileStream(savefilename.ToString().Split('.').GetValue(0).ToString()+"$.html",FileMode.Create);
streamHtmlHelp = new System.IO.StreamWriter(fs,System.Text.Encoding.GetEncoding("GB2312"));
streamHtmlHelp.WriteLine("<a href=\""+n.Text+"\">源檔案下載</a><br>");
do 
{
    str = objreader.ReadLine();
    streamHtmlHelp.WriteLine(str);
}
while (str !=  "</html>");
streamHtmlHelp.Close();
objreader.Close();
File.Delete(savefilename.ToString());
File.Move(savefilename.ToString().Split('.').GetValue(0).ToString()+"$.html",savefilename.ToString());

Excel -〉Html（這個與Word不同，因為Excel為很多層，又並列很多層，必須都清空才能銷毀執行個體，但實際中我發現並不是每次都能銷毀，所以網上求解多次沒有結果，只能殺沒進程，為了保證只殺滅最近的進程，我用時間進行判斷）

Microsoft.Office.Interop.Excel.Application   repExcel   =   new   Microsoft.Office.Interop.Excel.Application ();//執行個體化Excel
Microsoft.Office.Interop.Excel.Workbook   workbook   =   null;         
Microsoft.Office.Interop.Excel.Worksheet  worksheet   =   null;
workbook   =   repExcel.Application.Workbooks.Open(Application.StartupPath+@"\ccyt_chm\inform\"+n.FullPath,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing);//開啟檔案，n.FullPath是檔案路徑
worksheet = (Microsoft.Office.Interop.Excel.Worksheet)workbook.Worksheets[1];
object   htmlFile   =   Application.StartupPath+@"\ccyt_chm\inform\"+n.FullPath.Split('.').GetValue(0)+".html";   
object   ofmt   =   Microsoft.Office.Interop.Excel.XlFileFormat.xlHtml;   
workbook.SaveAs(htmlFile,ofmt,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Microsoft.Office.Interop.Excel.XlSaveAsAccessMode.xlNoChange,Type.Missing,Type.Missing,Type.Missing,Type.Missing,Type.Missing);//進行另存新檔操作   
object   osave   =   false;         
workbook.Close(osave,Type.Missing,Type.Missing);//逐步關閉所有使用的對象
repExcel.Quit();
System.Runtime.InteropServices.Marshal.ReleaseComObject(worksheet);
worksheet = null;
GC.Collect();//記憶體回收
System.Runtime.InteropServices.Marshal.ReleaseComObject(workbook);
workbook=null;
GC.Collect();
System.Runtime.InteropServices.Marshal.ReleaseComObject(repExcel.Application.Workbooks);
GC.Collect();
System.Runtime.InteropServices.Marshal.ReleaseComObject(repExcel);
repExcel = null;
GC.Collect();
System.Diagnostics.Process[] process = System.Diagnostics.Process.GetProcessesByName("EXCEL");//依據時間殺滅進程
foreach ( System.Diagnostics.Process p in process)
{
    if (DateTime.Now.Second - p.StartTime.Second > 0 && DateTime.Now.Second - p.StartTime.Second < 5)
    {
          p.Kill();
     }
}
Thread.Sleep(3000);//保證完全關閉
StreamReader objreader = new StreamReader(htmlFile.ToString().Split('.').GetValue(0)+".files\\sheet001.html",System.Text.Encoding.GetEncoding("GB2312"));//以下內容是在Html的第一個架構中添加下載原Excel的超連結         
FileStream fs = new FileStream(htmlFile.ToString().Split('.').GetValue(0)+".files\\sheet001$.html",FileMode.Create);
streamHtmlHelp = new System.IO.StreamWriter(fs,System.Text.Encoding.GetEncoding("GB2312"));
streamHtmlHelp.WriteLine("<a href=\""+"../"+n.Text+"\">源檔案下載</a><br>");
do 
{
    str = objreader.ReadLine();
    streamHtmlHelp.WriteLine(str);
}
while (str !=  "</html>");
streamHtmlHelp.Close();
objreader.Close();
File.Delete(htmlFile.ToString().Split('.').GetValue(0)+".files\\sheet001.html");
File.Move(htmlFile.ToString().Split('.').GetValue(0)+".files\\sheet001$.html",htmlFile.ToString().Split('.').GetValue(0)+".files\\sheet001.html");