基于Spark的电影推荐系统（推荐系统~2）

第四部分-推荐系统-数据ETL

• 本模块完成数据清洗，并将清洗后的数据load到Hive数据表里面去

前置准备：

spark +hive

vim $SPARK_HOME/conf/hive-site.xml
<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<configuration>
<property>
<name>hive.metastore.uris</name>
<value>thrift://hadoop001:9083</value>
</property>
</configuration>

• 启动Hive metastore server

[root@hadoop001 conf]# nohup hive --service metastore &

[root@hadoop001 conf]# netstat -tanp | grep 9083
tcp 0 0 0.0.0.0:9083 0.0.0.0:* LISTEN 24787/java
[root@hadoop001 conf]#

测试：
[root@hadoop001 ~]# spark-shell --master local[2]

scala> spark.sql("select * from liuge_db.dept").show;
+------+-------+-----+
|deptno|  dname|  loc|
+------+-------+-----+
|     1|  caiwu| 3lou|
|     2|  renli| 4lou|
|     3|  kaifa| 5lou|
|     4|qiantai| 1lou|
|     5|lingdao|4 lou|
+------+-------+-----+

==》保证Spark SQL 能够访问到Hive 的元数据才行。

然而我们采用的是standalone模式：需要启动master worker
[root@hadoop001 sbin]# pwd
/root/app/spark-2.4.3-bin-2.6.0-cdh5.7.0/sbin
[root@hadoop001 sbin]# ./start-all.sh

[root@hadoop001 sbin]# jps
26023 Master
26445 Worker

Spark常用端口

8080    spark.master.ui.port    Master WebUI
8081    spark.worker.ui.port    Worker WebUI
18080   spark.history.ui.port   History server WebUI
7077    SPARK_MASTER_PORT       Master port
6066    spark.master.rest.port  Master REST port
4040    spark.ui.port           Driver WebUI

这个时候打开：http://hadoop001:8080/

开始项目Coding

IDEA+Scala+Maven进行项目的构建

步骤一： 新建scala项目后，可以参照如下pom进行配置修改

<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/maven-v4_0_0.xsd">
<modelVersion>4.0.0</modelVersion>
<groupId>com.csylh</groupId>
<artifactId>movie-recommend</artifactId>
<version>1.0</version>
<inceptionYear>2008</inceptionYear>
<properties>
<scala.version>2.11.8</scala.version>
<spark.version>2.4.3</spark.version>
</properties>

<repositories>
<repository>
<id>scala-tools.org</id>
<name>Scala-Tools Maven2 Repository</name>
<url>http://scala-tools.org/repo-releases</url>
</repository>
</repositories>

<dependencies>

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-core_2.11</artifactId>
<version>${spark.version}</version>
</dependency>

<dependency>
<groupId>org.apache.hadoop</groupId>
<artifactId>hadoop-client</artifactId>
<version>2.6.0</version>
</dependency>

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-sql_2.11</artifactId>
<version>${spark.version}</version>
</dependency>

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-hive_2.11</artifactId>
<version>${spark.version}</version>
</dependency>

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-mllib_2.11</artifactId>
<version>${spark.version}</version>
</dependency>

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-streaming_2.11</artifactId>
<version>${spark.version}</version>
</dependency>

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-streaming-kafka-0-8_2.11</artifactId>
<version>${spark.version}</version>
</dependency>

<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka-clients</artifactId>
<version>1.1.1</version>
</dependency>
<!--// 0.10.2.1-->

<dependency>
<groupId>mysql</groupId>
<artifactId>mysql-connector-java</artifactId>
<version>5.1.39</version>
</dependency>

<dependency>
<groupId>log4j</groupId>
<artifactId>log4j</artifactId>
<version>1.2.17</version>
</dependency>

</dependencies>

<build>
<!--<sourceDirectory>src/main/scala</sourceDirectory>-->
<!--<testSourceDirectory>src/test/scala</testSourceDirectory>-->
<plugins>
<plugin>
<!-- see http://davidb.github.com/scala-maven-plugin -->
<groupId>net.alchim31.maven</groupId>
<artifactId>scala-maven-plugin</artifactId>
<version>3.1.3</version>
<executions>
<execution>
<goals>
<goal>compile</goal>
<goal>testCompile</goal>
</goals>
<configuration>
<args>
<arg>-dependencyfile</arg>
<arg>${project.build.directory}/.scala_dependencies</arg>
</args>
</configuration>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-surefire-plugin</artifactId>
<version>2.13</version>
<configuration>
<useFile>false</useFile>
<disableXmlReport>true</disableXmlReport>
<!-- If you have classpath issue like NoDefClassError,... -->
<!-- useManifestOnlyJar>false</useManifestOnlyJar -->
<includes>
<include>**/*Test.*</include>
<include>**/*Suite.*</include>
</includes>
</configuration>
</plugin>
</plugins>
</build>
</project>

步骤二：新建com.csylh.recommend.dataclearer.SourceDataETLApp

import com.csylh.recommend.entity.{Links, Movies, Ratings, Tags}
import org.apache.spark.sql.{SaveMode, SparkSession}

/**
* Description:
*    hadoop001  file:///root/data/ml/ml-latest 下的文件
*    ====>  SparkSQL ETL
*    ===>  load data to Hive数据仓库
*
* @Author: 留歌36
* @Date: 2019-07-12 13:48
*/
object SourceDataETLApp{
def main(args: Array[String]): Unit = {
// 面向SparkSession编程
val spark = SparkSession.builder()
//          .master("local[2]")
.enableHiveSupport() //开启访问Hive数据, 要将hive-site.xml等文件放入Spark的conf路径
.getOrCreate()

val sc = spark.sparkContext

// 设置RDD的partitions 的数量一般以集群分配给应用的CPU核数的整数倍为宜， 4核8G ，设置为8就可以
// 问题一：为什么设置为CPU核心数的整数倍？
// 问题二：数据倾斜，拿到数据大的partitions的处理，会消耗大量的时间，因此做数据预处理的时候，需要考量会不会发生数据倾斜
val minPartitions = 8

//  在生产环境中一定要注意设置spark.sql.shuffle.partitions，默认是200,及需要配置分区的数量
val shuffleMinPartitions = "8"
spark.sqlContext.setConf("spark.sql.shuffle.partitions",shuffleMinPartitions)
/**
* 1
*/
import spark.implicits._
val links = sc.textFile("file:///root/data/ml/ml-latest/links.txt",minPartitions) //DRIVER
.filter(!_.endsWith(",")) //EXRCUTER
.map(_.split(",")) //EXRCUTER
.map(x => Links(x(0).trim.toInt, x(1).trim.toInt, x(2).trim.toInt)) //EXRCUTER
.toDF()
println("===============links===================:",links.count())
links.show()

// 把数据写入到HDFS上
links.write.mode(SaveMode.Overwrite).parquet("/tmp/links")

// 将数据从HDFS加载到Hive数据仓库中去
spark.sql("drop table if exists links")
spark.sql("create table if not exists links(movieId int,imdbId int,tmdbId int) stored as parquet")
spark.sql("load data inpath '/tmp/links' overwrite into table links")

/**
* 2
*/
val movies = sc.textFile("file:///root/data/ml/ml-latest/movies.txt",minPartit
2d96
ions)
.filter(!_.endsWith(","))
.map(_.split(","))
.map(x => Movies(x(0).trim.toInt, x(1).trim.toString, x(2).trim.toString))
.toDF()
println("===============movies===================:",movies.count())
movies.show()

// 把数据写入到HDFS上
movies.write.mode(SaveMode.Overwrite).parquet("/tmp/movies")

// 将数据从HDFS加载到Hive数据仓库中去
spark.sql("drop table if exists movies")
spark.sql("create table if not exists movies(movieId int,title String,genres String) stored as parquet")
spark.sql("load data inpath '/tmp/movies' overwrite into table movies")

/**
* 3
*/
val ratings = sc.textFile("file:///root/data/ml/ml-latest/ratings.txt",minPartitions)
.filter(!_.endsWith(","))
.map(_.split(","))
.map(x => Ratings(x(0).trim.toInt, x(1).trim.toInt, x(2).trim.toDouble, x(3).trim.toInt))
.toDF()
println("===============ratings===================:",ratings.count())

ratings.show()

// 把数据写入到HDFS上
ratings.write.mode(SaveMode.Overwrite).parquet("/tmp/ratings")

// 将数据从HDFS加载到Hive数据仓库中去
spark.sql("drop table if exists ratings")
spark.sql("create table if not exists ratings(userId int,movieId int,rating Double,timestamp int) stored as parquet")
spark.sql("load data inpath '/tmp/ratings' overwrite into table ratings")

/**
* 4
*/
val tags = sc.textFile("file:///root/data/ml/ml-latest/tags.txt",minPartitions)
.filter(!_.endsWith(","))
.map(x => rebuild(x))  // 注意这个坑的解决思路
.map(_.split(","))
.map(x => Tags(x(0).trim.toInt, x(1).trim.toInt, x(2).trim.toString, x(3).trim.toInt))
.toDF()

tags.show()

// 把数据写入到HDFS上
tags.write.mode(SaveMode.Overwrite).parquet("/tmp/tags")

// 将数据从HDFS加载到Hive数据仓库中去
spark.sql("drop table if exists tags")
spark.sql("create table if not exists tags(userId int,movieId int,tag String,timestamp int) stored as parquet")
spark.sql("load data inpath '/tmp/tags' overwrite into table tags")
}
/**
* 该方法是用于处理不符合规范的数据
* @param input
* @return
*/
private def rebuild(input:String): String ={
val a = input.split(",")

val head = a.take(2).mkString(",")
val tail = a.takeRight(1).mkString
val tag = a.drop(2).dropRight(1).mkString.replaceAll("\"","")
val output = head + "," + tag + "," + tail
output
}
}

再有一些上面主类引用到的case 对象，你可以理解为Java 实体类

package com.csylh.recommend.entity

/**
* Description: 数据的schema
*
* @Author: 留歌36
* @Date: 2019-07-12 13:46
*/
case class Links(movieId:Int,imdbId:Int,tmdbId:Int)

package com.csylh.recommend.entity

/**
* Description: TODO
*
* @Author: 留歌36
* @Date: 2019-07-12 14:09
*/
case class Movies(movieId:Int,title:String,genres:String)

package com.csylh.recommend.entity

/**
* Description: TODO
*
* @Author: 留歌36
* @Date: 2019-07-12 14:10
*/
case class Ratings(userId:Int,movieId:Int,rating:Double,timestamp:Int)

package com.csylh.recommend.entity

/**
* Description: TODO
*
* @Author: 留歌36
* @Date: 2019-07-12 14:11
*/
case class Tags(userId:Int,movieId:Int,tag:String,timestamp:Int)

步骤三：将创建的项目进行打包上传到服务器
mvn clean package -Dmaven.test.skip=true

[root@hadoop001 ml]# ll -h movie-recommend-1.0.jar
-rw-r--r--. 1 root root 156K 10月 20 13:56 movie-recommend-1.0.jar
[root@hadoop001 ml]#

步骤四：提交运行上面的jar,编写shell脚本

[root@hadoop001 ml]# vim etl.sh
export HADOOP_CONF_DIR=/root/app/hadoop-2.6.0-cdh5.7.0/etc/hadoop

$SPARK_HOME/bin/spark-submit --class com.csylh.recommend.dataclearer.SourceDataETLApp --master spark://hadoop001:7077 --name SourceDataETLApp --driver-memory 10g --executor-memory 5g /root/data/ml/movie-recommend-1.0.jar

步骤五：sh etl.sh 即可

先把数据写入到HDFS上
创建Hive表
load 数据到表

sh etl.sh之前：

[root@hadoop001 ml]# hadoop fs -ls /tmp
19/10/20 19:26:58 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Found 2 items
drwx------   - root supergroup          0 2019-04-01 16:27 /tmp/hadoop-yarn
drwx-wx-wx   - root supergroup          0 2019-04-02 09:33 /tmp/hive

[root@hadoop001 ml]# hadoop fs -ls /user/hive/warehouse
19/10/20 19:27:03 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
[root@hadoop001 ml]#

sh etl.sh之后：
这里的shell 是 ,spark on standalone,后面会spark on yarn。其实也没差，都可以

[root@hadoop001 ~]# hadoop fs -ls /tmp
19/10/20 19:43:17 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Found 6 items
drwx------   - root supergroup          0 2019-04-01 16:27 /tmp/hadoop-yarn
drwx-wx-wx   - root supergroup          0 2019-04-02 09:33 /tmp/hive
drwxr-xr-x   - root supergroup          0 2019-10-20 19:42 /tmp/links
drwxr-xr-x   - root supergroup          0 2019-10-20 19:42 /tmp/movies
drwxr-xr-x   - root supergroup          0 2019-10-20 19:43 /tmp/ratings
drwxr-xr-x   - root supergroup          0 2019-10-20 19:43 /tmp/tags
[root@hadoop001 ~]# hadoop fs -ls /user/hive/warehouse
19/10/20 19:43:32 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Found 4 items
drwxr-xr-x   - root supergroup          0 2019-10-20 19:42 /user/hive/warehouse/links
drwxr-xr-x   - root supergroup          0 2019-10-20 19:42 /user/hive/warehouse/movies
drwxr-xr-x   - root supergroup          0 2019-10-20 19:43 /user/hive/warehouse/ratings
drwxr-xr-x   - root supergroup          0 2019-10-20 19:43 /user/hive/warehouse/tags
[root@hadoop001 ~]#

这样我们就把数据etl到我们的数据仓库里了，接下来，基于这份基础数据做数据加工

有任何问题，欢迎留言一起交流~~
更多文章：基于Spark的电影推荐系统：https://blog.csdn.net/liuge36/column/info/29285