4.MapReduce 序列化-编程知识

概述

序列化是分布式计算中很重要的一环境，好的序列化方式，可以大大减少分布式计算中，网络传输的数据量。

序列化

对象 --> 字节序例：存储到磁盘或者网络传输
MR 、Spark、Flink ：分布式的执行框架必然会涉及到网络传输

java 中的序列化：Serializable
Hadoop 中序列化特点：紧凑、速度、扩展性、互操作
Spark 中使用了其它的序例化框架 Kyro

反序例化

字节序例 —> 对象

java自带的两种

Serializable

此处是 java 自带的 序例化 方式，这种方式简单方便，但体积大，不利于大数据量网络传输。

public class JavaSerDemo {public static void main(String[] args) throws IOException, ClassNotFoundException {Person person = new Person(1, "张三", 33);ObjectOutputStream out = new ObjectOutputStream(new FileOutputStream("download/person.obj"));out.writeObject(person);ObjectInputStream in = new ObjectInputStream(new FileInputStream("download/person.obj"));Object o = in.readObject();System.out.println(o);}static class Person implements Serializable {private int id;private String name;private int age;public Person(int id, String name, int age) {this.id = id;this.name = name;this.age = age;}@Overridepublic String toString() {return "Person{" +"id=" + id +", name='" + name + '\'' +", age=" + age +'}';}public int getId() {return id;}public void setId(int id) {this.id = id;}public String getName() {return name;}public void setName(String name) {this.name = name;}public int getAge() {return age;}public void setAge(int age) {this.age = age;}}
}

非Serializable

public class DataSerDemo {public static void main(String[] args) throws IOException {Person person = new Person(1, "张三", 33);DataOutputStream out = new DataOutputStream(new FileOutputStream("download/person2.obj"));out.writeInt(person.getId());out.writeUTF(person.getName());out.close();DataInputStream in = new DataInputStream(new FileInputStream("download/person2.obj"));// 这里要注意，上面以什么顺序写出去，这里就要以什么顺序读取int id = in.readInt();String name = in.readUTF();in.close();System.out.println("id:" + id + " name:" + name);}/***  注意： 不需要继承 Serializable*/static class Person {private int id;private String name;private int age;public Person(int id, String name, int age) {this.id = id;this.name = name;this.age = age;}@Overridepublic String toString() {return "Person{" +"id=" + id +", name='" + name + '\'' +", age=" + age +'}';}public int getId() {return id;}public void setId(int id) {this.id = id;}public String getName() {return name;}public void setName(String name) {this.name = name;}public int getAge() {return age;}public void setAge(int age) {this.age = age;}}
}

hadoop序例化

官方地址速递

The key and value classes have to be serializable by the framework and hence need to implement the Writable interface. Additionally, the key classes have to implement the WritableComparable interface to facilitate sorting by the framework.

注意：Writable 两个方法，一个 write ,readFields

@InterfaceAudience.Public
@InterfaceStability.Stable
public interface Writable {void write(DataOutput out) throws IOException;void readFields(DataInput in) throws IOException;
}

实践

public class PersonWritable implements Writable {private int id;private String name;private int age;// 消费金额private int consumption;// 消费总金额private long consumptions;public PersonWritable() {}public PersonWritable(int id, String name, int age, int consumption) {this.id = id;this.name = name;this.age = age;this.consumption = consumption;}public PersonWritable(int id, String name, int age, int consumption, long consumptions) {this.id = id;this.name = name;this.age = age;this.consumption = consumption;this.consumptions = consumptions;}public int getId() {return id;}public void setId(int id) {this.id = id;}public String getName() {return name;}public void setName(String name) {this.name = name;}public int getAge() {return age;}public void setAge(int age) {this.age = age;}public int getConsumption() {return consumption;}public void setConsumption(int consumption) {this.consumption = consumption;}public long getConsumptions() {return consumptions;}public void setConsumptions(long consumptions) {this.consumptions = consumptions;}@Overridepublic String toString() {return"id=" + id +", name='" + name + '\'' +", age='" + age + '\'' +", consumption=" + consumption + '\'' +", consumptions=" + consumptions;}@Overridepublic void write(DataOutput out) throws IOException {out.writeInt(id);out.writeUTF(name);out.writeInt(age);out.writeInt(consumption);out.writeLong(consumptions);}@Overridepublic void readFields(DataInput in) throws IOException {id = in.readInt();name = in.readUTF();age = in.readInt();consumption = in.readInt();consumptions = in.readLong();}
}

/*** 统计 个人 消费*/
public class PersonStatistics {static class PersonStatisticsMapper extends Mapper<LongWritable, Text, IntWritable, PersonWritable> {@Overrideprotected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {String[] split = value.toString().split(",");int id = Integer.parseInt(split[0]);String name = split[1];int age = Integer.parseInt(split[2]);int consumption = Integer.parseInt(split[3]);PersonWritable writable = new PersonWritable(id, name, age, consumption, 0);context.write(new IntWritable(id), writable);}}static class PersonStatisticsReducer extends Reducer<IntWritable, PersonWritable, NullWritable, PersonWritable> {@Overrideprotected void reduce(IntWritable key, Iterable<PersonWritable> values, Context context) throws IOException, InterruptedException {long count = 0L;PersonWritable person = null;for (PersonWritable data : values) {if (Objects.isNull(person)) {person = data;}count = count + data.getConsumption();}person.setConsumptions(count);PersonWritable personWritable = new PersonWritable(person.getId(), person.getName(), person.getAge(), person.getConsumption(), count);context.write(NullWritable.get(), personWritable);}}public static void main(String[] args) throws IOException, InterruptedException, ClassNotFoundException {Configuration configuration = new Configuration();String sourcePath = "data/person.data";String distPath = "downloadOut/person-out.data";FileUtil.deleteIfExist(configuration, distPath);Job job = Job.getInstance(configuration, "person statistics");job.setJarByClass(PersonStatistics.class);//job.setCombinerClass(PersonStatistics.PersonStatisticsReducer.class);job.setMapperClass(PersonStatisticsMapper.class);job.setReducerClass(PersonStatisticsReducer.class);job.setMapOutputKeyClass(IntWritable.class);job.setMapOutputValueClass(PersonWritable.class);job.setOutputKeyClass(NullWritable.class);job.setOutputValueClass(PersonWritable.class);FileInputFormat.addInputPath(job, new Path(sourcePath));FileOutputFormat.setOutputPath(job, new Path(distPath));System.exit(job.waitForCompletion(true) ? 0 : 1);}
}

# person.data
1,张三,30,10
1,张三,30,20
2,李四,25,5

上述执行结果如下：
在这里插入图片描述

分片/InputFormat & InputSplit

官方文档速递

org.apache.hadoop.mapreduce.InputFormat
org.apache.hadoop.mapreduce.InputSplit

日志

执行 序列化 测试小程序，关注以下日志

# 总共加载一个文件，分隔成一个
2024-01-06 09:19:42,363 [main] [org.apache.hadoop.mapreduce.lib.input.FileInputFormat] [INFO] - Total input files to process : 1
2024-01-06 09:19:42,487 [main] [org.apache.hadoop.mapreduce.JobSubmitter] [INFO] - number of splits:1