几种读取文件方式的性能测试

来源：互联网发布：北京华为海思媒体算法编辑：程序博客网时间：2024/05/22 14:41

这些天一直在对文件进行读写操作，为了性能，换了好几种方式，我下面贴出来：

/*** 随机读取文件内容* * @param fileName*/public static String readFileByRandomAccess(String path) {StringBuffer sb = new StringBuffer();RandomAccessFile randomFile = null;try {// 打开一个随机访问文件流，按只读方式randomFile = new RandomAccessFile(path, "rw");byte[] bytes = new byte[1024];int byteread = 0;while ((byteread = randomFile.read(bytes)) != -1) {// System.out.write(bytes, 0, byteread);sb.append(new String(bytes, 0, byteread));}} catch (IOException e) {e.printStackTrace();} finally {if (randomFile != null) {try {randomFile.close();} catch (IOException e1) {}}}return sb.toString();}

/*** 以行为单位读取文件，常用于读面向行的格式化文件* * @param fileName*/public static String readFileByLines(String path) {StringBuffer sb = new StringBuffer();BufferedReader reader = null;try {reader = new BufferedReader(new FileReader(path));String tempString = null;while ((tempString = reader.readLine()) != null) {sb.append(tempString).append(System.getProperty("line.separator"));// 显示行号}} catch (IOException e) {e.printStackTrace();} finally {try {if (reader != null) {reader.close();}} catch (IOException e) {System.err.println("流关闭失败");}}return sb.toString();}

/*** 使用Java.nio ByteBuffer字节将一个文件输出* * @param filePath*/public static String readFileByBybeBuffer(String path) {StringBuffer sb = new StringBuffer();FileInputStream in = null;FileOutputStream out = null;try {// 获取源文件和目标文件的输入输出流in = new FileInputStream(path);// 获取输入输出通道FileChannel fcIn = in.getChannel();ByteBuffer buffer = ByteBuffer.allocate(1024);CharBuffer charBuffer = null;Charset charset = Charset.forName("UTF-8");CharsetDecoder decoder = charset.newDecoder();while (true) {// clear方法重设缓冲区，使它可以接受读入的数据buffer.clear();// 从输入通道中将数据读到缓冲区int r = fcIn.read(buffer);if (r == -1) {break;}charBuffer = decoder.decode(buffer);// flip方法让缓冲区可以将新读入的数据写入另一个通道buffer.flip();sb.append(charBuffer.toString());}} catch (Exception e) {e.printStackTrace();} finally {try {if (in != null && out != null) {in.close();out.close();}} catch (IOException e) {e.printStackTrace();}}//System.out.println(sb.toString());return sb.toString();}

/*** 通过scanner的方式读取文件* * @param fileName* @return*/public static String readByScanner(String path) {StringBuffer sb = new StringBuffer();Scanner s = null;try {File f = new File(path);s = new Scanner(f);while (s.hasNext()) {sb.append(s.next());}} catch (FileNotFoundException e) {e.printStackTrace();} finally {if(s!=null)s.close();}return sb.toString();}

我们来测试下：

String fileName="Z:\\flume\\flume\\conf\\ErrLog";long t1=OperateUtil.getTime();OperateUtil.readByScanner(fileName);long t2=OperateUtil.getTime();OperateUtil.readFileByBybeBuffer(fileName);long t3=OperateUtil.getTime();OperateUtil.readFileByLines(fileName);long t4=OperateUtil.getTime();OperateUtil.readFileByRandomAccess(fileName);long t5=OperateUtil.getTime();System.out.println("readByScanner:"+(t2-t1));System.out.println("readFileByBybeBuffer:"+(t3-t2));System.out.println("readFileByLines:"+(t4-t3));System.out.println("readFileByRandomAccess:"+(t5-t4));ErrLog的文件大小是65M。

下面看下测试结果：

readByScanner:6180readFileByBybeBuffer:171readFileByLines:468readFileByRandomAccess:531

经过多次测试，依旧发现readFileByBybeBuffer这个方式的效率更好，使用的是NIO

0 0