MR实现reduce join和map join及hive的执行计划
一、涵盖
MapReduce
InputFormat
RecordReader
切片：block=input split 1.1
File…
Text…
NLine…
DB…
Mapper
setup
map 业务逻辑
cleanup
Combiner
本地的Reducer
注意适用场景
Partitioner
将key按照某种规则进行分发
Hash：
Custom
Reducer
setup
reduce
cleanup
OutputFormat
RecordWriter
DB…
Text…
File…
排序
全排
分区排
二、MapReduce Hive实现join
join
select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno;
Reduce/Shuffle/Common Join
join操作是在reduce端完成的
mr:
两张表 emp dept
使用两个MapTask来读取我们的emp和dept数据
MR是要进行shuffle，Mapper的key是我们on的条件 deptno
相同的deptno会被分发到同一个reduce中去
有一个标识符来区分某一条数据到底来自于哪个表
select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno;
最终的输出字段：empno ename deptno dname flag
需要自定义序列化类：Info
脏数据 不干净的数据
文件：MapTask
表：TableScan
1）读完数据以后==>Shuffle==>Reduce来join
2）某一个join的key对应的数据量大==>skew
1Y条数据
9000W都是deptno=10 ReduceTask
reducer join
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0modelVersion>
<groupId>com.cc.pxj.wfygroupId>
<artifactId>phoneWcRuoZeartifactId>
<version>1.0-SNAPSHOTversion>
<properties>
<project.build.sourceEncoding>UTF-8project.build.sourceEncoding>
<maven.compiler.source>1.8maven.compiler.source>
<maven.compiler.target>1.8maven.compiler.target>
<hadoop.version>2.6.0-cdh5.16.2hadoop.version>
properties>
<repositories>
<repository>
<id>clouderaid>
<url>https://repository.cloudera.com/artifactory/cloudera-repos/url>
repository>
repositories>
<dependencies>
<dependency>
<groupId>org.apache.hadoopgroupId>
<artifactId>hadoop-clientartifactId>
<version>${hadoop.version}version>
dependency>
<dependency>
<groupId>junitgroupId>
<artifactId>junitartifactId>
<version>4.11version>
<scope>testscope>
dependency>
<dependency>
<groupId>mysqlgroupId>
<artifactId>mysql-connector-javaartifactId>
<version>5.1.17version>
dependency>
dependencies>
<build>
<pluginManagement>
<plugins>
<plugin>
<artifactId>maven-clean-pluginartifactId>
<version>3.1.0version>
plugin>
<plugin>
<artifactId>maven-resources-pluginartifactId>
<version>3.0.2version>
plugin>
<plugin>
<artifactId>maven-compiler-pluginartifactId>
<version>3.8.0version>
plugin>
<plugin>
<artifactId>maven-surefire-pluginartifactId>
<version>2.22.1version>
plugin>
<plugin>
<artifactId>maven-jar-pluginartifactId>
<version>3.0.2version>
plugin>
<plugin>
<artifactId>maven-install-pluginartifactId>
<version>2.5.2version>
plugin>
<plugin>
<artifactId>maven-deploy-pluginartifactId>
<version>2.8.2version>
plugin>
<plugin>
<artifactId>maven-site-pluginartifactId>
<version>3.7.1version>
plugin>
<plugin>
<artifactId>maven-project-info-reports-pluginartifactId>
<version>3.0.0version>
plugin>
plugins>
pluginManagement>
build>
project>
package com.ccj.pxj.join;
import org.apache.hadoop.io.Writable;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
/**
*
*/
public class Info implements Writable {
private int empno;
private String ename;
private int deptno;
private String dname;
private int flag; //区分数据从哪里来的标志位
public Info(){}
public Info(int empno, String ename, int deptno, String dname){
this.empno = empno;
this.ename = ename;
this.deptno = deptno;
this.dname = dname;
}
public Info(int empno, String ename, int deptno, String dname, int flag){
this(empno, ename,deptno,dname);
this.flag = flag;
}
@Override
public String toString() {
return
empno + "\t" + ename + "\t" + deptno +
"\t" + dname ;
}
@Override
public void write(DataOutput out) throws IOException {
out.writeInt(empno);
out.writeUTF(ename);
out.writeInt(deptno);
out.writeUTF(dname);
out.writeInt(flag);
}
@Override
public void readFields(DataInput in) throws IOException {
this.empno = in.readInt();
this.ename = in.readUTF();
this.deptno = in.readInt();
this.dname = in.readUTF();
this.flag = in.readInt();
}
public int getEmpno() {
return empno;
}
public void setEmpno(int empno) {
this.empno = empno;
}
public String getEname() {
return ename;
}
public void setEname(String ename) {
this.ename = ename;
}
public int getDeptno() {
return deptno;
}
public void setDeptno(int deptno) {
this.deptno = deptno;
}
public String getDname() {
return dname;
}
public void setDname(String dname) {
this.dname = dname;
}
public int getFlag() {
return flag;
}
public void setFlag(int flag) {
this.flag = flag;
}
}
package com.ccj.pxj.join;
import com.ccj.pxj.phone.utils.FileUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.FileSplit;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
/**
*
*/
public class ReduceJoinDriver {
public static void main(String[] args) throws Exception{
String input = "data/join";
String output = "outreducejoin";
Configuration configuration = new Configuration();
Job job = Job.getInstance(configuration);
FileUtils.deleteOutput(configuration, output);
job.setJarByClass(ReduceJoinDriver.class);
job.setMapperClass(MyMapper.class);
job.setReducerClass(MyReducer.class);
job.setMapOutputKeyClass(IntWritable.class);
job.setMapOutputValueClass(Info.class);
job.setOutputKeyClass(Info.class);
job.setOutputValueClass(NullWritable.class);
FileInputFormat.setInputPaths(job, new Path(input));
FileOutputFormat.setOutputPath(job, new Path(output));
boolean result = job.waitForCompletion(true);
System.exit(result ? 0 : 1);
}
/**
* 在Mapper中如何区分哪个数据是哪个表过来的....
*/
public static class MyMapper extends Mapper<LongWritable, Text, IntWritable, Info> {
String name; // 标识数据从哪来的
@Override
protected void setup(Context context) throws IOException, InterruptedException {
FileSplit split = (FileSplit) context.getInputSplit();
name = split.getPath().getName();
}
@Override
protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
String[] splits = value.toString().split("\t");
if (name.contains("emp")) { // 处理emp的数据
if(splits.length == 8) {
// 1
int empno = Integer.parseInt(splits[0].trim());
String ename = splits[1];
int deptno = Integer.parseInt(splits[7].trim());
Info info = new Info(empno, ename, deptno, "", 1); // 1:emp
context.write(new IntWritable(deptno), info);
}
} else { // dept
int deptno = Integer.parseInt(splits[0].trim());
Info info = new Info(0, "", deptno, splits[1], 2); // 2:dept
context.write(new IntWritable(deptno), info);
}
}
}
// 真正的join操作是在此处完成的
public static class MyReducer extends Reducer<IntWritable, Info, Info, NullWritable> {
@Override
protected void reduce(IntWritable key, Iterable<Info> values, Context context) throws IOException, InterruptedException {
List<Info> infos = new ArrayList<>();
String dname = "";
for (Info info : values) {
if (info.getFlag() == 1) { // emp
Info tmp = new Info();
tmp.setEmpno(info.getEmpno());
tmp.setEname(info.getEname());
tmp.setDeptno(info.getDeptno());
infos.add(tmp);
} else if (info.getFlag() == 2) { // dept
dname = info.getDname();
}
// System.out.println(infos.size()+".......");
}
for (Info bean : infos) {
bean.setDname(dname);
context.write(bean, NullWritable.get());
}
}
}
}
数据：
dept:
10 ACCOUNTING NEW YORK
20 RESEARCH DALLAS
30 SALES CHICAGO
40 OPERATIONS BOSTON
emp
7369 SMITH CLERK 7902 1980-12-17 800.00 20
7499 ALLEN SALESMAN 7698 1981-2-20 1600.00 300.00 30
7521 WARD SALESMAN 7698 1981-2-22 1250.00 500.00 30
7566 JONES MANAGER 7839 1981-4-2 2975.00 20
7654 MARTIN SALESMAN 7698 1981-9-28 1250.00 1400.00 30
7698 BLAKE MANAGER 7839 1981-5-1 2850.00 30
7782 CLARK MANAGER 7839 1981-6-9 2450.00 10
7788 SCOTT ANALYST 7566 1987-4-19 3000.00 20
7839 KING PRESIDENT 1981-11-17 5000.00 10
7844 TURNER SALESMAN 7698 1981-9-8 1500.00 0.00 30
7876 ADAMS CLERK 7788 1987-5-23 1100.00 20
7900 JAMES CLERK 7698 1981-12-3 950.00 30
7902 FORD ANALYST 7566 1981-12-3 3000.00 20
7934 MILLER CLERK 7782 1982-1-23 1300.00 10
8888 HIVE PROGRAM 7839 1988-1-23 10300.00
map join
package com.ccj.pxj.join;
import org.apache.hadoop.io.Writable;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
/**
*
*/
public class Info implements Writable {
private int empno;
private String ename;
private int deptno;
private String dname;
private int flag; //区分数据从哪里来的标志位
public Info(){}
public Info(int empno, String ename, int deptno, String dname){
this.empno = empno;
this.ename = ename;
this.deptno = deptno;
this.dname = dname;
}
public Info(int empno, String ename, int deptno, String dname, int flag){
this(empno, ename,deptno,dname);
this.flag = flag;
}
@Override
public String toString() {
return
empno + "\t" + ename + "\t" + deptno +
"\t" + dname ;
}
@Override
public void write(DataOutput out) throws IOException {
out.writeInt(empno);
out.writeUTF(ename);
out.writeInt(deptno);
out.writeUTF(dname);
out.writeInt(flag);
}
@Override
public void readFields(DataInput in) throws IOException {
this.empno = in.readInt();
this.ename = in.readUTF();
this.deptno = in.readInt();
this.dname = in.readUTF();
this.flag = in.readInt();
}
public int getEmpno() {
return empno;
}
public void setEmpno(int empno) {
this.empno = empno;
}
public String getEname() {
return ename;
}
public void setEname(String ename) {
this.ename = ename;
}
public int getDeptno() {
return deptno;
}
public void setDeptno(int deptno) {
this.deptno = deptno;
}
public String getDname() {
return dname;
}
public void setDname(String dname) {
this.dname = dname;
}
public int getFlag() {
return flag;
}
public void setFlag(int flag) {
this.flag = flag;
}
}
package com.ccj.pxj.join;
import com.ccj.pxj.phone.utils.FileUtils;
import org.apache.commons.lang.StringUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.*;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.FileSplit;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import java.io.BufferedReader;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStreamReader;
import java.net.URI;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
/**
*
*/
public class MapJoinDriver {
public static void main(String[] args) throws Exception {
String input = "data/join/emp.txt";
String output = "out";
Configuration configuration = new Configuration();
Job job = Job.getInstance(configuration);
FileUtils.deleteOutput(configuration, output);
job.setJarByClass(MapJoinDriver.class);
job.setMapperClass(MyMapper.class);
job.setNumReduceTasks(0); // 不需要reduce
job.addCacheFile(new URI("data/join/dept.txt"));
job.setOutputKeyClass(Info.class);
job.setOutputValueClass(NullWritable.class);
FileInputFormat.setInputPaths(job, new Path(input));
FileOutputFormat.setOutputPath(job, new Path(output));
boolean result = job.waitForCompletion(true);
System.exit(result ? 0 : 1);
}
/**
* 在Mapper中如何区分哪个数据是哪个表过来的....
*/
public static class MyMapper extends Mapper<LongWritable, Text, IntWritable, Info> {
Map<String, String> cache = new HashMap<>();
@Override
protected void setup(Context context) throws IOException, InterruptedException {
String path = context.getCacheFiles()[0].getPath().toString();
BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream(path)));
String line;
while (StringUtils.isNotEmpty(line = reader.readLine())) {
String[] splits = line.split("\t");
cache.put(splits[0], splits[1]); // 部门编号和部门名称
}
IOUtils.closeStream(reader);
}
@Override
protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
String[] splits = value.toString().split("\t");
if (splits.length == 8) {
int empno = Integer.parseInt(splits[0].trim());
String ename = splits[1];
int deptno = Integer.parseInt(splits[7].trim());
Info info = new Info(empno, ename, deptno, cache.get(deptno+""));
context.write(new IntWritable(deptno), info);
}
}
}
}
造数据脚本
package com.ccj.wfy.makedata;
import java.io.*;
import java.util.Random;
public class Mock {
public static void main(String[] args) throws Exception {
BufferedWriter writer = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(new File("data/wc.txt"))));
String[] a={"pxj","wfy","ccj","zmj","xwc","lzh","zcl","wlp","wxc","pk","jpeson"};
for (int i=0;i<100;i++){
Random random = new Random();
writer.write(a[random.nextInt(a.length)]);
writer.write(",");
writer.write(a[random.nextInt(a.length)]);
writer.write(",");
writer.write(a[random.nextInt(a.length)]);
writer.newLine();
}
writer.flush();
writer.close();
}
}
三、执行计划
建表
CREATE TABLE `dept`(
`deptno` int,
`dname` string,
`loc` string)
row format delimited
fields terminated by '\t'
CREATE TABLE `emp`(
`empno` int,
`ename` string,
`job` string,
`mgr` int,
`hiredate` string,
`sal` double,
`comm` double,
`deptno` int)
row format delimited
fields terminated by '\t'
hive (default)> load data local inpath '//home/pxj/datas/dept.txt' into table dept;
hive (default)> load data local inpath '/home/pxj/datas/emp.txt' into table emp;
先关闭
set hive.auto.convert.join
0: jdbc:hive2://pxj:10000> set hive.auto.convert.join=false
. . . . . . . . . . . . .> ;
0: jdbc:hive2://pxj:10000> explain select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno;
INFO : Compiling command(queryId=pxj_20200119010707_26c4bbd5-479b-49f1-8968-b08a1abc5056): explain select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno
INFO : Semantic Analysis Completed
INFO : Returning Hive schema: Schema(fieldSchemas:[FieldSchema(name:Explain, type:string, comment:null)], properties:null)
INFO : Completed compiling command(queryId=pxj_20200119010707_26c4bbd5-479b-49f1-8968-b08a1abc5056); Time taken: 0.243 seconds
INFO : Concurrency mode is disabled, not creating a lock manager
INFO : Executing command(queryId=pxj_20200119010707_26c4bbd5-479b-49f1-8968-b08a1abc5056): explain select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno
INFO : Starting task [Stage-3:EXPLAIN] in serial mode
INFO : Completed executing command(queryId=pxj_20200119010707_26c4bbd5-479b-49f1-8968-b08a1abc5056); Time taken: 0.042 seconds
INFO : OK
+----------------------------------------------------+--+
| Explain |
+----------------------------------------------------+--+
| STAGE DEPENDENCIES: |
| Stage-1 is a root stage |
| Stage-0 depends on stages: Stage-1 |
| |
| STAGE PLANS: |
| Stage: Stage-1 |
| Map Reduce |
| Map Operator Tree: |
| TableScan |
| alias: e |
| Statistics: Num rows: 6 Data size: 700 Basic stats: COMPLETE Column stats: NONE |
| Filter Operator |
| predicate: deptno is not null (type: boolean) |
| Statistics: Num rows: 3 Data size: 350 Basic stats: COMPLETE Column stats: NONE |
| Reduce Output Operator |
| key expressions: deptno (type: int) |
| sort order: + |
| Map-reduce partition columns: deptno (type: int) |
| Statistics: Num rows: 3 Data size: 350 Basic stats: COMPLETE Column stats: NONE |
| value expressions: empno (type: int), ename (type: string) |
| TableScan |
| alias: d |
| Statistics: Num rows: 1 Data size: 79 Basic stats: COMPLETE Column stats: NONE |
| Filter Operator |
| predicate: deptno is not null (type: boolean) |
| Statistics: Num rows: 1 Data size: 79 Basic stats: COMPLETE Column stats: NONE |
| Reduce Output Operator |
| key expressions: deptno (type: int) |
| sort order: + |
| Map-reduce partition columns: deptno (type: int) |
| Statistics: Num rows: 1 Data size: 79 Basic stats: COMPLETE Column stats: NONE |
| value expressions: dname (type: string) |
| Reduce Operator Tree: |
| Join Operator |
| condition map: |
| Inner Join 0 to 1 |
| keys: |
| 0 deptno (type: int) |
| 1 deptno (type: int) |
| outputColumnNames: _col0, _col1, _col7, _col12 |
| Statistics: Num rows: 3 Data size: 385 Basic stats: COMPLETE Column stats: NONE |
| Select Operator |
| expressions: _col0 (type: int), _col1 (type: string), _col7 (type: int), _col12 (type: string) |
| outputColumnNames: _col0, _col1, _col2, _col3 |
| Statistics: Num rows: 3 Data size: 385 Basic stats: COMPLETE Column stats: NONE |
| File Output Operator |
| compressed: false |
| Statistics: Num rows: 3 Data size: 385 Basic stats: COMPLETE Column stats: NONE |
| table: |
| input format: org.apache.hadoop.mapred.TextInputFormat |
| output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat |
| serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe |
| |
| Stage: Stage-0 |
| Fetch Operator |
| limit: -1 |
| Processor Tree: |
| ListSink |
| |
+----------------------------------------------------+--+
59 rows selected (0.379 seconds)
说明：
执行计划
explain sql
一个sql一次性能完成吗？不太一定
一个hivesql可能会被拆分成多个mr作业
ABSTRACT SYNTAX TREE：AST
STAGE DEPENDENCIES
STAGE PLANS
number of mappers: 2; number of reducers: 1
数据倾斜：Shuffle
mapjoin执行计划
0: jdbc:hive2://pxj:10000> set hive.auto.convert.join=true;
No rows affected (0.004 seconds)
0: jdbc:hive2://pxj:10000> explain select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno;
INFO : Compiling command(queryId=pxj_20200119011010_f0ca7412-1b4e-41bf-9c1c-34c6d801acff): explain select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno
INFO : Semantic Analysis Completed
INFO : Returning Hive schema: Schema(fieldSchemas:[FieldSchema(name:Explain, type:string, comment:null)], properties:null)
INFO : Completed compiling command(queryId=pxj_20200119011010_f0ca7412-1b4e-41bf-9c1c-34c6d801acff); Time taken: 0.197 seconds
INFO : Concurrency mode is disabled, not creating a lock manager
INFO : Executing command(queryId=pxj_20200119011010_f0ca7412-1b4e-41bf-9c1c-34c6d801acff): explain select e.empno,e.ename,e.deptno,d.dname from emp e join dept d on e.deptno=d.deptno
INFO : Starting task [Stage-5:EXPLAIN] in serial mode
INFO : Completed executing command(queryId=pxj_20200119011010_f0ca7412-1b4e-41bf-9c1c-34c6d801acff); Time taken: 0.015 seconds
INFO : OK
+----------------------------------------------------+--+
| Explain |
+----------------------------------------------------+--+
| STAGE DEPENDENCIES: |
| Stage-4 is a root stage |
| Stage-3 depends on stages: Stage-4 |
| Stage-0 depends on stages: Stage-3 |
| |
| STAGE PLANS: |
| Stage: Stage-4 |
| Map Reduce Local Work |
| Alias -> Map Local Tables: |
| d |
| Fetch Operator |
| limit: -1 |
| Alias -> Map Local Operator Tree: |
| d |
| TableScan |
| alias: d |
| Statistics: Num rows: 1 Data size: 79 Basic stats: COMPLETE Column stats: NONE |
| Filter Operator |
| predicate: deptno is not null (type: boolean) |
| Statistics: Num rows: 1 Data size: 79 Basic stats: COMPLETE Column stats: NONE |
| HashTable Sink Operator |
| keys: |
| 0 deptno (type: int) |
| 1 deptno (type: int) |
| |
| Stage: Stage-3 |
| Map Reduce |
| Map Operator Tree: |
| TableScan |
| alias: e |
| Statistics: Num rows: 6 Data size: 700 Basic stats: COMPLETE Column stats: NONE |
| Filter Operator |
| predicate: deptno is not null (type: boolean) |
| Statistics: Num rows: 3 Data size: 350 Basic stats: COMPLETE Column stats: NONE |
| Map Join Operator |
| condition map: |
| Inner Join 0 to 1 |
| keys: |
| 0 deptno (type: int) |
| 1 deptno (type: int) |
| outputColumnNames: _col0, _col1, _col7, _col12 |
| Statistics: Num rows: 3 Data size: 385 Basic stats: COMPLETE Column stats: NONE |
| Select Operator |
| expressions: _col0 (type: int), _col1 (type: string), _col7 (type: int), _col12 (type: string) |
| outputColumnNames: _col0, _col1, _col2, _col3 |
| Statistics: Num rows: 3 Data size: 385 Basic stats: COMPLETE Column stats: NONE |
| File Output Operator |
| compressed: false |
| Statistics: Num rows: 3 Data size: 385 Basic stats: COMPLETE Column stats: NONE |
| table: |
| input format: org.apache.hadoop.mapred.TextInputFormat |
| output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat |
| serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe |
| Local Work: |
| Map Reduce Local Work |
| |
| Stage: Stage-0 |
| Fetch Operator |
| limit: -1 |
| Processor Tree: |
| ListSink |
| |
+----------------------------------------------------+--+
62 rows selected (0.303 seconds)
说明：
Map Join
join操作是在map端完成的
有Shuffle吗？有没有Reduce？
思路：小表加载到内存中，读取大表时，读一条和内存中的数据匹配一下，匹配上就表示join上，没匹配88
缺点：没办法处理大数据量的表 适合大表join小表
mr不太适合交互次数比较多的场景
DAG RDD Tez
Hive: MR Tez Spark
以WC为例分析数据倾斜
同key分发到同一个reduce去执行==>skew
数据倾斜解决方案：打散再处理
pxj_8 100
pxj_1 200
pxj_2 300
pxj_4 200
==> 原来都在一个reducetask中去处理的数据，被我们打散的
9000w/100 这种处理方式 压力是不是很多了？
pxj 800
chain:
第一个mr：key加随机数打散
pxj_8 100
pxj_1 200
pxj_2 300
pxj_4 200
第二个mr：第一个mr加的随机数去掉
pxj 100
pxj 200
pxj300
pxj 200
在第二个mr中的reduce来做最终的聚合操作
pxj 800
木桶原理：水桶能装多少水，取决于你的水桶最短的那块板
MR：100reducetask，你整个作业的运行时长必然是以最后一个task执行完算的
一个task就把整个mr作业的执行性能拉低了
Hadoop不怕数据量大，就怕数据倾斜
数据热度/数据热点
作者：pxj（潘陈）