几种读取文件方式的性能测试

╰+攻爆jí腚メ 2023-10-16 21:36 37阅读 0赞

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:6180
    readFileByBybeBuffer:171
    readFileByLines:468
    readFileByRandomAccess: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