0x00 文章内容
- 编写代码
- 测试结果
0x01 编写代码
前提：因为需要用到Hadoop，所以需要先引入Hadoop相关的jar包
<dependency> <groupId>org.apache.hadoop</groupId> <artifactId>hadoop-client</artifactId> <version>2.7.5</version> </dependency>
1. 编写对象类
a. 编写Block类
package com.shaonaiyi.hadoop.serialize; import org.apache.hadoop.io.Writable; import java.io.DataInput; import java.io.DataOutput; import java.io.IOException; /** * @Author shaonaiyi@163.com * @Date 2019/12/13 16:51 * @Description 定义引入Hadoop序列化机制的BlockWritable类 */ public class BlockWritable implements Writable { private long blockId; private long numBytes; private long generationStamp; public BlockWritable(long blockId) { this.blockId = blockId; } public BlockWritable(long blockId, long numBytes, long generationStamp) { this.blockId = blockId; this.numBytes = numBytes; this.generationStamp = generationStamp; } public long getBlockId() { return blockId; } public long getNumBytes() { return numBytes; } public long getGenerationStamp() { return generationStamp; } @Override public String toString() { return "Block{" + "blockId=" + blockId + ", numBytes=" + numBytes + ", generationStamp=" + generationStamp + '}'; } public void write(DataOutput dataOutput) throws IOException { dataOutput.writeLong(blockId); dataOutput.writeLong(numBytes); dataOutput.writeLong(generationStamp); } public void readFields(DataInput dataInput) throws IOException { this.blockId = dataInput.readLong(); this.numBytes = dataInput.readLong(); this.generationStamp = dataInput.readLong(); } }
2. 编写调用测试代码
a. 编写序列化与反序列化代码
package com.shaonaiyi.hadoop.serialize; import org.apache.hadoop.io.IntWritable; import org.apache.hadoop.io.Text; import org.apache.hadoop.io.Writable; import org.apache.hadoop.io.WritableFactories; import java.io.*; /** * @Author shaonaiyi@163.com * @Date 2019/12/13 16:53 * @Description 编写Hadoop序列化机制测试类代码 */ public class HadoopSerializableTest { public static void main(String[] args) throws IOException { String fileName = "blockWritable.txt"; serialize(fileName); // deSerialize(fileName); } private static void serialize(String fileName) throws IOException { BlockWritable block = new BlockWritable(78062621L, 39447651L, 56737546L); File file = new File(fileName); if (file.exists()) { file.delete(); } file.createNewFile(); FileOutputStream fileOutputStream = new FileOutputStream(file); DataOutputStream dataOutputStream = new DataOutputStream(fileOutputStream); block.write(dataOutputStream); dataOutputStream.close(); } private static void deSerialize(String fileName) throws IOException { FileInputStream fileInputStream = new FileInputStream(fileName); DataInputStream dataInputStream = new DataInputStream(fileInputStream); Writable writable = WritableFactories.newInstance(BlockWritable.class); writable.readFields(dataInputStream); System.out.println((BlockWritable)writable); } }
0x02 测试结果
1. 测试序列化
a. 执行序列化方法，发现项目与main文件夹同级目录下多了一个blockWritable.txt文件，双击打开：
查看大小，发现是24字节：
我们定义的BlockWritable类，Long类是8个字节，三个Long类型的属性是24个字节，没有引入额外的信息，是我们想要的结果，请与上一篇教程做比较：Java的序列化与反序列化实操。
2. 测试反序列化
a. 打开反序列化方法，执行：
// serialize(fileName); deSerialize(fileName);
发现可以将BlockWritable对象反序列化出来了。
3. 测试修改对象代码
a. 尝试修改BlockWritable类，如在里面添加一个无参数的构造方法
public BlockWritable() { }
b. 然后再反序列化，发现执行的结果并没有报错，与前面一样。
0xFF 总结
- Hadoop的序列化机制解决了内置的Java序列化接口的缺陷。
- 本教程拥有前置教程：Java的序列化与反序列化实操！
- Hadoop的序列化机制接口如下：
在平时代码中，可以参考如下：
Text text = new Text(); String word = "hello"; text.set(word); IntWritable intWritable = new IntWritable(3);
