1.数据
这里只插入部分数据,有一些实验的数据中的日期格式是1998-XX-XX形式,在后来代码也会有相应的修改
13 987 1/10/1998 3 999 1 1232.16
13 1660 1/10/1998 3 999 1 1232.16
13 1762 1/10/1998 3 999 1 1232.16
13 1843 1/10/1998 3 999 1 1232.16
13 1948 1/10/1998 3 999 1 1232.16
13 2273 1/10/1998 3 999 1 1232.16
13 2380 1/10/1998 3 999 1 1232.16
13 2683 1/10/1998 3 999 1 1232.16
13 2865 1/10/1998 3 999 1 1232.16
13 4663 1/10/1998 3 999 1 1232.16
13 5203 1/10/1998 3 999 1 1232.16
13 5321 1/10/1998 3 999 1 1232.16
13 5590 1/10/1998 3 999 1 1232.16
13 6277 1/10/1998 3 999 1 1232.16
13 6859 1/10/1998 3 999 1 1232.16
13 8540 1/10/1998 3 999 1 1232.16
13 9076 1/10/1998 3 999 1 1232.16
13 12099 1/10/1998 3 999 1 1232.16
13 35834 1/10/1998 3 999 1 1232.16
13 524 1/20/1998 2 999 1 1205.99
13 188 1/20/1998 3 999 1 1232.16
13 361 1/20/1998 3 999 1 1232.16
13 531 1/20/1998 3 999 1 1232.16
13 659 1/20/1998 3 999 1 1232.16
13 848 1/20/1998 3 999 1 1232.16
13 949 1/20/1998 3 999 1 1232.16
13 1242 1/20/1998 3 999 1 1232.16
13 1291 1/20/1998 3 999 1 1232.16
13 1422 1/20/1998 3 999 1 1232.16
13 1485 1/20/1998 3 999 1 1232.16
13 1580 1/20/1998 3 999 1 1232.16
13 1943 1/20/1998 3 999 1 1232.16
13 1959 1/20/1998 3 999 1 1232.16
13 2021 1/20/1998 3 999 1 1232.16
13 2142 1/20/1998 3 999 1 1232.16
13 3014 1/20/1998 3 999 1 1232.16
13 3053 1/20/1998 3 999 1 1232.16
13 3261 1/20/1998 3 999 1 1232.16
13 3783 1/20/1998 3 999 1 1232.16
13 3947 1/20/1998 3 999 1 1232.16
13 4523 1/20/1998 3 999 1 1232.16
13 5813 1/20/1998 3 999 1 1232.16
13 6543 1/20/1998 3 999 1 1232.16
13 7076 1/20/1998 3 999 1 1232.16
13 7421 1/20/1998 3 999 1 1232.16
13 8747 1/20/1998 3 999 1 1232.16
13 8787 1/20/1998 3 999 1 1232.16
13 8836 1/20/1998 3 999 1 1232.16
13 9052 1/20/1998 3 999 1 1232.16
13 9680 1/20/1998 3 999 1 1232.16
2.要求
编写程序,将数据表中的销售笔数相加并求出各年的销售总额
原理
其实这里的原理和前面的wordcount的原理是一样的,只不过求的东西发生了变化罢了
3.代码
关于pom.xml文件的配置,在前面的文章已经有介绍—>这里
sold类
import org.apache.hadoop.io.Writable;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
public class sold implements Writable {
//字段名 prod_id,cust_id,time,channel_id,promo_id,quantity_sold,amount_sold
//数据类型:Int,Int,Date, Int,Int ,Int ,float(10,2),
//数据: 13, 987, 1998/1/10, 3, 999,1, 1232.16
//由以上定义变量
private int prod_id;
private int cust_id;
private String time;
private int channel_id;
private int promo_id;
private int quantity_sold;
private float amount_sold;
//序列化方法:将java对象转化为可跨机器传输数据流(二进制串/字节)的一种技术
public void write(DataOutput out) throws IOException {
out.writeInt(this.prod_id);
out.writeInt(this.cust_id);
out.writeUTF(this.time);
out.writeInt(this.channel_id);
out.writeInt(this.promo_id);
out.writeInt(this.quantity_sold);
out.writeFloat(this.amount_sold);
}
//反序列化方法:将可跨机器传输数据流(二进制串)转化为java对象的一种技术
public void readFields(DataInput in) throws IOException {
this.prod_id = in.readInt();
this.cust_id = in.readInt();
this.time = in.readUTF();
this.channel_id = in.readInt();
this.promo_id = in.readInt();
this.quantity_sold = in.readInt();
this.amount_sold = in.readFloat();
}
public int getProd_id() {
return prod_id;
}
public void setProd_id(int prod_id) {
this.prod_id = prod_id;
}
public int getCust_id() {
return cust_id;
}
public void setCust_id(int cust_id) {
this.cust_id = cust_id;
}
public String getTime() {
return time;
}
public void setTime(String time) {
this.time = time;
}
public int getChannel_id() {
return channel_id;
}
public void setChannel_id(int channel_id) {
this.channel_id = channel_id;
}
public int getPromo_id() {
return promo_id;
}
public void setPromo_id(int promo_id) {
this.promo_id = promo_id;
}
public int getQuantity_sold() {
return quantity_sold;
}
public void setQuantity_sold(int quantity_sold) {
this.quantity_sold = quantity_sold;
}
public float getAmount_sold() {
return amount_sold;
}
public void setAmount_sold(float amount_sold) {
this.amount_sold = amount_sold;
}
}
sold_mapper第一类
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;
import java.io.IOException;
public class sold_mapper extends Mapper<LongWritable, Text, Text, sold> {
@Override
protected void map(LongWritable k1, Text v1, Context context) throws IOException, InterruptedException {
String data = v1.toString();
String[] words = data.split(",");
sold s = new sold();
s.setTime("1998");
s.setQuantity_sold(Integer.parseInt(words[5]));
s.setAmount_sold(Float.valueOf(words[6]));
context.write(new Text("1998"), s);
}
}
对于csv文件中日期格式为1998-XX-XX的文件而言,mapper类需要修改一下
sold_mapper第二类
import org.apache.commons.lang.StringUtils;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;
import java.io.IOException;
public class SoldMapper extends Mapper<LongWritable, Text, Text, Sold> {
@Override
protected void map(LongWritable k1, Text v1,
Context context)
throws IOException, InterruptedException {
//字段名 prod_id,cust_id,time,channel_id,promo_id,quantity_sold,amount_sold
//数据: 13,987,1998-01-10,3,999,1,1232.16
String data = v1.toString();
String[] words = data.split(",");
//数据: t1=987,1998-01-10,3,999,1,1232.16
String t1 = StringUtils.substringAfter(data, ",");
//数据: t2=1998-01-10,3,999,1,1232.16
String t2 = StringUtils.substringAfter(t1, ",");
//取年份为偏移量,数据: words2[0]=1998,words2[1]=01,words2[2]=10,3,999,1,1232.16
String[] words2 = t2.split("-");
// StringUtils.substringAfter("dskeabcedeh", "e");
// /*结果是:abcedeh*/
Sold sold = new Sold();
sold.setTime(words[2]);//数组word[]
sold.setQuantity_sold(Integer.parseInt(words[5]));
sold.setAmount_sold(Float.valueOf(words[6]));
context.write(new Text(words2[0]), sold);//数组word2[],word2[0]代表年份作为k2
}
}
sold_reducer类
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Reducer;
import java.io.IOException;
public class sold_reducer extends Reducer<Text, sold,Text,Text> {
protected void reduce(Text k3,Iterable<sold> v3,Context context) throws IOException,InterruptedException {
int sum1 = 0;
float sum2 = 0;
for(sold s:v3){
sum1 = sum1 + s.getQuantity_sold();
sum2 = sum2 + s.getAmount_sold();
}
String end = "销售数量: "+Integer.toString(sum1)+", "+"销售总额: "+Float.toString(sum2);
context.write(k3,new Text(end));
}
}
sold_main类
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
public class sold_main {
public static void main(String[] args) throws Exception {
//1. 创建一个job和任务入口(指定主类)
Job job = Job.getInstance(new Configuration());
job.setJarByClass(sold_main.class);
//2. 指定job的mapper和输出的类型<k2 v2>
job.setMapperClass(sold_mapper.class);
job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(sold.class);
//3. 指定job的reducer和输出的类型<k4 v4>
job.setReducerClass(sold_reducer.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(Text.class);
//4. 指定job的输入和输出路径
FileInputFormat.setInputPaths(job, new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));
//5. 执行job
job.waitForCompletion(true);
}
}
4.打包jar包
之后我们需要右击项目找到Open Module Settings
找到Artifacts→JAR→From modules with dependencies…
找到主类Myemployee_partitioner
在主页面找到Build → Build Artifacts→Build
之后我们可以发现目录树中出现了一个out目录,在字树文件中有一个
mypro.jar
文件,这个就是我们打包的文件
5.将打包的jar文件上传到hdfs
hdfs dfs -put 虚拟机本地jar包存储完整路径 hdfs文件系统存储的路径
6.利用Hadoop集群运行该jar包
hadoop jar 项目名.jar 主类名 输入目录 输出目录
eg:hadoop jar sold.jar sold_main /input/sales.csv /output
7.演示
在打包完jar包之后,开始对terminal写入命令
查看输出结果与实验要求是否一致
由于我的文件是只有1998年这一年的数据,所以输出结果中只有这一年的销售数量和销售总额
8.总结
其实我们通过wordcount就已经知道了mapreduce的整体流程,对于后来的其他要求只是对于map函数和reduce函数的修改罢了,要记住,MapTask的数量是有split决定的,而split的大小默认是一个block的大小,我们可以通过修改split的大小来控制,而map输出后的中间数据集中会以key来决定ReduceTask的数量。