基于IntelliJ IDEA开发Spark的Maven项目——Scala语言

转自：http://blog.csdn.net/kwu_ganymede/article/details/51832427

基于IntelliJ IDEA开发Spark的Maven项目——Scala语言

1、Maven管理项目在JavaEE普遍使用，开发Spark项目也不例外，而scala语言开发Spark项目的首选。因此需要构建Maven-Scala项目来开发Spark项目，本文采用的工具是IntelliJ
IDEA 2016，IDEA工具越来越被大家认可，开发Java, Python ,scala
支持都非常好

下载链接 ： https://www.jetbrains.com/idea/download/

安装直接下一步即可

2、安装scala插件，File->Settings->Editor->Plugins,搜索scala即可安装



可能由于网络的原因下载不了，可以采取离线安装的方式，例如：



提示下载失败后，根据提示的地址下载离线安装包 http://plugins.jetbrains.com/files/631/24825/python-145.86.zip

在界面选择离线安装即可：



3、创建Maven工程，File->New Project->Maven

选择相应的JDK版本，直接下一步



设定Maven项目的GroupId及ArifactId



创建项目的工程名称，点击完成即可


创建Maven工程完毕，默认是Java的，没关系后面我们再添加scala与spark的依赖



4、修改Maven项目的pom.xml文件，增加scala与spark的依赖

[java] view
plain copy

<?xml version="1.0" encoding="UTF-8"?>  

<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>4.0.0</modelVersion>  

  

    <groupId>com.ganymede</groupId>  

    <artifactId>sparkplatformstudy</artifactId>  

    <version>1.0-SNAPSHOT</version>  

  

    <properties>  

        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>  

        <spark.version>1.6.0</spark.version>  

        <scala.version>2.10</scala.version>  

        <hadoop.version>2.6.0</hadoop.version>  

    </properties>  

  

    <dependencies>  

        <dependency>  

            <groupId>org.apache.spark</groupId>  

            <artifactId>spark-core_${scala.version}</artifactId>  

            <version>${spark.version}</version>  

        </dependency>  

        <dependency>  

            <groupId>org.apache.spark</groupId>  

            <artifactId>spark-sql_${scala.version}</artifactId>  

            <version>${spark.version}</version>  

        </dependency>  

        <dependency>  

            <groupId>org.apache.spark</groupId>  

            <artifactId>spark-hive_${scala.version}</artifactId>  

            <version>${spark.version}</version>  

        </dependency>  

        <dependency>  

            <groupId>org.apache.spark</groupId>  

            <artifactId>spark-streaming_${scala.version}</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-streaming-kafka_${scala.version}</artifactId>  

            <version>${spark.version}</version>  

        </dependency>  

        <dependency>  

            <groupId>org.apache.spark</groupId>  

            <artifactId>spark-mllib_${scala.version}</artifactId>  

            <version>${spark.version}</version>  

        </dependency>  

        <dependency>  

            <groupId>mysql</groupId>  

            <artifactId>mysql-connector-java</artifactId>  

            <version>5.1.39</version>  

        </dependency>  

        <dependency>  

            <groupId>junit</groupId>  

            <artifactId>junit</artifactId>  

            <version>4.12</version>  

        </dependency>  

    </dependencies>  

  

    <!-- maven官方 http://repo1.maven.org/maven2/  或 http://repo2.maven.org/maven2/ （延迟低一些） -->  

    <repositories>  

        <repository>  

            <id>central</id>  

            <name>Maven Repository Switchboard</name>  

            <layout>default</layout>  

            <url>http://repo2.maven.org/maven2</url>  

            <snapshots>  

                <enabled>false</enabled>  

            </snapshots>  

        </repository>  

    </repositories>  

  

    <build>  

        <sourceDirectory>src/main/scala</sourceDirectory>  

        <testSourceDirectory>src/test/scala</testSourceDirectory>  

  

        <plugins>  

            <plugin>  

                <!-- MAVEN 编译使用的JDK版本 -->  

                <groupId>org.apache.maven.plugins</groupId>  

                <artifactId>maven-compiler-plugin</artifactId>  

                <version>3.3</version>  

                <configuration>  

                    <source>1.7</source>  

                    <target>1.7</target>  

                    <encoding>UTF-8</encoding>  

                </configuration>  

            </plugin>  

        </plugins>  

    </build>  

</project>  

5、删除项目的java目录，新建scala并设置源文件夹



添加scala的SDK



添加scala的SDK成功



6、开发Spark实例



测试案例来自spark官网的mllib例子 http://spark.apache.org/docs/latest/mllib-data-types.html

[java] view
plain copy

import org.apache.spark.{SparkConf, SparkContext}  

  

/** 

  * Created by wuke on 2016/7/5. 

  */  

object LoadLibSVMFile extends App{  

  import org.apache.spark.mllib.regression.LabeledPoint  

  import org.apache.spark.mllib.util.MLUtils  

  import org.apache.spark.rdd.RDD  

  

  val conf = new SparkConf().setAppName("LogisticRegressionMail").setMaster("local")  

  

  val sc = new SparkContext(conf)  

  val examples: RDD[LabeledPoint] = MLUtils.loadLibSVMFile(sc, "data/mllib/sample_libsvm_data.txt")  

  

  println(examples.first)  

}  

测试通过

7、打包编译，线上发布



注意选择依赖包