使用C#实现将Word 转文本存储到数据库并进行管理

CREATE TABLE [dbo].[f_words](
    [cid] [uniqueidentifier] ROWGUIDCOL  NOT NULL,
    [filename] [nvarchar](100) NOT NULL,
    [bfile] [image] NULL,
    [fcontent] [nvarchar](max) NULL,
    [sys_instime] [datetime] NULL,
 CONSTRAINT [PK_f_words] PRIMARY KEY CLUSTERED 
(
    [cid] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]
GO
 
ALTER TABLE [dbo].[f_words] ADD  CONSTRAINT [DF_f_words_cid]  DEFAULT (newid()) FOR [cid]
GO

public string getWordTxt(string _filename,bool getHtmlContent)
        {
            resultReport = "";
            Object Nothing = System.Reflection.Missing.Value;
 
            object filename = _filename;
            //创建一个名为WordApp的组件对象
            DateTime beforetime = DateTime.Now;
            Word.Application WordApp = new Word.Application();
            //创建一个名为WordDoc的文档对象
            WordApp.DisplayAlerts = Word.WdAlertLevel.wdAlertsNone;
 
            Word.Document WordDoc = WordApp.Documents.Open(ref filename, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing, ref Nothing);
 
            WordDoc.SpellingChecked = false;//关闭拼写检查
 
            WordDoc.ShowSpellingErrors = false;//关闭显示拼写错误提示框
 
            DateTime aftertime = DateTime.Now;
 
            string rv = WordDoc.Content.Text;
            Sys_Custom_DocVar = "";
            Sys_Custom_DocVar2 = "";
            foreach (Word.Variable ov in WordDoc.Variables)
            {
                if (ov.Name == "sys_custom_docvar")
                {
                    //                    WordDoc.Content.Text = ov.Value;
                    Sys_Custom_DocVar = ov.Value;
                } else if (ov.Name == "sys_custom_docvar2")
                {
                    //                    WordDoc.Content.Text = ov.Value;
                    Sys_Custom_DocVar2 = ov.Value;
                }
            }
            foreach (Word.ContentControl cc in WordDoc.ContentControls)
            {
                resultReport += cc.ID + ":" + cc.Tag + "<br>";
 
            }
            string _path = Path.GetDirectoryName(_filename) + "\\";
 
            object _expFile = _path + Guid.NewGuid().ToString() + ".html";
            if (getHtmlContent == true)
            {
                object   wsf = Microsoft.Office.Interop.Word.WdSaveFormat.wdFormatHTML;
                WordDoc.SaveAs2(ref _expFile,ref wsf, ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing,ref Nothing);
            }
            WordDoc.Close(ref Nothing, ref Nothing, ref Nothing);
            //关闭WordApp组件对象
            WordApp.Quit(ref Nothing, ref Nothing, ref Nothing);
 
            KillProcessByStartTime("WINWORD",beforetime,aftertime);
 
            if (File.Exists(_expFile.ToString()) == true)
            {
                FileEx fe = new FileEx();
                rv = fe.LoadFromFile(_expFile.ToString(), Encoding.Default);
                File.Delete(_expFile.ToString());
            }
            return rv;
        }
 
public string KillProcessByStartTime(string processName,DateTime beforetime,DateTime aftertime)
        {
            Process[] ps = Process.GetProcesses();
            foreach (Process p in ps)  
            {
                if(p.ProcessName.ToUpper()!=processName) continue;
                if(p.StartTime > beforetime && p.StartTime < aftertime)
                {
                    try
                    {
                        p.Kill();
                    }
                    catch(Exception e)
                    {
                        return e.Message;
                    }
                }
            }  
            return "";
        }

string filename = Request.PhysicalApplicationPath + "\\app_data\\" + Guid.NewGuid().ToString() + ".docx";  //预生成文件名
//File1为上传控件
File1.PostedFile.SaveAs(filename);  //保存文件
 
//添加SQL参数，此处仅为示例
ArrayList paras = new ArrayList();
paras.Add(new SqlParameter("filename", filename));
paras.Add(new SqlParameter("fcontent", getWordTxt(filename,false)));  //word转文本
paras.Add(new SqlParameter("bfile", GetBinaryData(filename)));  //word的二进制信息
paras.Add(new SqlParameter("sys_instime", System.DateTime.Now));
 
File.Delete(filename);
 
//保存到数据表
ExecDbScripts("INSERT INTO [f_words]([filename],[bfile],[fcontent],[sys_instime])  VALUES(@filename, @bfile,@fcontent,@sys_instime)", paras);
得到文件Byte[]数据方法
public byte[] GetBinaryData(string filename)
{
    if(!File.Exists(filename))
    {
        return null;
    }
    FileStream fs = new FileStream(filename, FileMode.Open, FileAccess.Read);
    byte[] imageData = new Byte[fs.Length];
    fs.Read( imageData, 0,Convert.ToInt32(fs.Length));
    fs.Close();
    return imageData;
}        

 
  string strConn =ConfigurationSettings.AppSettings["Connection"];
  SqlConnection Conn = new SqlConnection(strConn );
  SqlCommand Cmd = new SqlCommand();
 
  Cmd.Connection = Conn;
 
  SqlDataReader myDr;
 
  Cmd.CommandText = " select filename from f_words where  cid=@cid ";
   
 
  SqlParameter   para2=new   SqlParameter("@cid",SqlDbType.UniqueIdentifier);
  para2.Value=(new Guid(_cid));
  Cmd.Parameters.Add(para2);
  try
  {
             Conn.Open();
             myDr = Cmd.ExecuteReader();
             bool _hasrows=myDr.HasRows;
             if (myDr.Read())
             {
 
                 string extendname = "docx";
 
                 byte[] bytes = (byte[])myDr["bfile"];
                 
                 
                 Response.Buffer = true;
                 Response.Charset = "utf-8";
                 Response.AppendHeader("Content-Disposition", "inline;filename=" + HttpUtility.UrlEncode(myDr["filename"].ToString() + "" + extendname)); //把 attachment 改为 online 则在线打开
                 Response.ContentEncoding = System.Text.Encoding.GetEncoding("utf-8");
                 Response.AppendHeader("Content-Length", bytes.Length.ToString());
                 Response.ContentType = "application/octet-stream"; 
                 
                     Page.EnableViewState = false;
                     Response.BinaryWrite(bytes);
                     Response.Flush();
             }
             myDr.Close();
         }
         catch (SqlException ex)
         {
         }
         finally
         {
             Conn.Close();
             Conn.Dispose();
         }
}