将word文档转化成字符串或者txt文档、将txt文档转化成字符串

=================================================================

/**

*

*判断是word文档还是txt文档

*/

 public String readFile() throws Exception {
   String fileRealName = "我们的幸福";   //文件名
  String fileString = null;
  if (null != fileRealName && !"".equals(fileRealName)) {
   try {
    File file = new File("d:\\test\\law\\" + fileRealName);   //d:\\test\\law\\是文档存放路径
    if (fileRealName.endsWith(".doc")) {
     fileString = WordReader.readDoc("d:\\test\\law\\"+ fileRealName);
    } else if (fileRealName.endsWith(".txt")) {
     fileString = readFile(file);
    }
   } catch (Exception e) {
    e.printStackTrace();
   }
  }
    return "moblieinfo";
 }

=================================================================

 

/**

*

*读取txt文档并返回字符串

*/

 private static String readFile(File file) throws FileNotFoundException,
   IOException {

  StringBuffer sb = new StringBuffer();
  FileReader fl = new FileReader(file);
  BufferedReader bf = new BufferedReader(fl);
  String context = null;
  do {
   context = bf.readLine();
   if (context == null) {
    break;
   } else if (!context.trim().equals("")) {
    sb.append(context + " ");
   }
  } while (context != null);
  bf.close();
  return sb.toString();
 }

 

======================================================================

package com.sdcncsi.common.util;

import java.io.File;
import java.io.FileInputStream;

import org.apache.poi.hwpf.extractor.WordExtractor;

public class WordReader {

 public WordReader() {

 }

 public static String readDoc(String doc) throws Exception {
  //       
  // 创建输入流读取DOC文件
  File newfile = new File(doc);
  System.out.println("--------------");
  FileInputStream in = new FileInputStream(newfile);
  WordExtractor extractor = null;
  String text = null;

  extractor = new WordExtractor(in);
  text = extractor.getText();

  in.close();

  return text;
 }

}
========================================================================

 

import java.io.*;
import org.apache.poi.hwpf.extractor.WordExtractor;

public class convertWordToTxt {
 public static String convertWordToTxt(FileInputStream fis,
   String outputPath, String outputFileName) {
  try {
   WordExtractor extractor = new WordExtractor(fis);
   String str = extractor.getText().trim(); // 获取Word文件中的字符

   String outputFile = outputPath + outputFileName + ".txt"; // 组装输出TXT文件的绝对路径

   FileOutputStream fos = new FileOutputStream(outputFile);
   fos.write(str.getBytes());
   fos.close();
  } catch (FileNotFoundException e) {
   return "文件不存在!";
  } catch (IOException e) {
   return "文件读写错误!";
  }
  return "OK";
 }
}
=======================================================================

备注：需要的jar包