在Spark中RDD转换为DataFrame一共有两种方法。一种是基于一个封装数据的类去转换，另一种是通过结构类型去转换，下面分别介绍。
基于封装数据的类去转换（反射方式）
Java版本
这种方法是底层基于类的反射机制去实现的，为了封装数据我们首先需要一个封装数据的类也就是JavaBean。然后去构造一个JavaRDD<JavaBean>，然后调用方法去构造DataFrame，这时需要传入构造好的RDD与JavaBean.class进而得到一个Dataset<Row>（注意在Java中Dataset也就是DataFrame）。下面做一个展示：
我们所用到的数据：
| 1,leo,17 2,marry,17 3,jack,18 4,tom,19 |
下面为Java代码：
首先构造Student的JavaBean
package test.scala.com.zhuliu7.sparksqlnext_java;import java.io.Serializable;public class Student implements Serializable {private Integer id;private String name;private Integer age;public void setId(Integer id) {this.id = id;}public void setName(String name) {this.name = name;}public void setAge(Integer age) {this.age = age;}public Integer getId() {return id;}public String getName() {return name;}public Integer getAge() {return age;}@Overridepublic String toString() {return "Student{" +"id=" + id +", name='" + name + '\'' +", age=" + age +'}';}
}
然后是正式的逻辑代码
package test.scala.com.zhuliu7.sparksqlnext_java;import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.api.java.function.VoidFunction;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Row;
import org.apache.spark.sql.SQLContext;public class RDD2DataFrameReflection {public static void main(String[] args) {SparkConf conf = new SparkConf().setAppName("dfuse").setMaster("local[*]");JavaSparkContext sc = new JavaSparkContext(conf);SQLContext sqlContext = new SQLContext(sc);JavaRDD<String> oraginRDD = sc.textFile("E:\\testdata\\students.txt");/** 在Java中如果我们知道RDD中的数据对应着一个JavaBean，这时要将RDD转换DF的话首先需要一个JavaBean类型的RDD* */JavaRDD<Student> stuRDD = oraginRDD.map(new Function<String, Student>() {@Overridepublic Student call(String s) throws Exception {String[] strs = s.split(",");Student stu = new Student();stu.setId(Integer.valueOf(strs[0]));stu.setName(strs[1]);stu.setAge(Integer.valueOf(strs[2]));return stu;}});// 使用反射的方式来创建DataFrame，包含的数据类型只能是简单类型，这句代码是核心Dataset<Row> ds = sqlContext.createDataFrame(stuRDD,Student.class);ds.show();ds.registerTempTable("t_stu");Dataset nds = sqlContext.sql("select * from t_stu where age <= 18");// 将查询出来的结果再次封装到RDD中JavaRDD<Row> finalRDD = nds.javaRDD();// 将数据映射为Student类型JavaRDD<Student> fRDD = finalRDD.map(new Function<Row, Student>() {@Overridepublic Student call(Row row) throws Exception {Student stu = new Student();stu.setAge(row.getInt(0));stu.setName(row.getString(2));stu.setId(row.getInt(1));return stu;}});fRDD.foreach(new VoidFunction<Student>() {@Overridepublic void call(Student student) throws Exception {System.out.println(student);}});}
}
从上面我们也能够看到，想要从DataFrame还原为以前的JavaRDD<JavaBean>只需要将DataFrame转换为JavaRDD<Row>再做一次映射进而转换为JavaRDD<JavaBean>就ok了。
Scala版本
与Java版本的思路一样只不过是把JavaBean替换成了case class，还需要注意的事项在以下注释中写出。代码如下：
import org.apache.spark.{SparkConf, SparkContext}
import org.apache.spark.sql.SQLContext
object RDD2DataFrameReflaction {case class Student(id : Int, name : String,age : Int)def main(args: Array[String]): Unit = {val conf = new SparkConf().setAppName("DFUse").setMaster("local[*]")val sc = new SparkContext(conf)val sqlContext = new SQLContext(sc)import sqlContext.implicits._ // 放在sqlContext之后val preRDD = sc.textFile("E:\\testdata\\students.txt").map(x => x.split(",")).map(arr => Student(arr(0).toInt,arr(1).toString,arr(2).toInt))val df = preRDD.toDF() // 这句话编译通过需要import sqlContext.implicits._的支持，而这句话能够正确运行需要将case class放在成员变量的位置df.show()// 接下来将df中查询的结果在还原为RDDdf.registerTempTable("t_sdu")val fdf = sqlContext.sql("select * from t_stu where age > 17")val tranRDD = fdf.rdd // 这时RDD中的元素都为Rowval finalRDD = tranRDD.map(row => Student(row(0).toString.toInt,row(1).toString,row(1).toString.toInt)) //这个顺序可以自定义val fRDD = tranRDD.map(row => Student(row.getInt(0),row.getString(1),row.getInt(2)))}}
使用结构类型StructType去转换
Java版本
首先去构造一个JavaRDD<Row>，这里要注意Row中的元素数据类型要和我们的即将构造的结构类型中的类型要一一对应。接着就要去构造StructType类型，最后使用SQLContext配合JavaRDD与StructType去创建DataFrame，代码如下：
package test.scala.com.zhuliu7.sparksqlnext_java;import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Row;
import org.apache.spark.sql.RowFactory;
import org.apache.spark.sql.SQLContext;
import org.apache.spark.sql.types.DataTypes;
import org.apache.spark.sql.types.StructField;
import org.apache.spark.sql.types.StructType;import java.util.ArrayList;
import java.util.List;public class Programme2DF {public static void main(String[] args) {SparkConf conf = new SparkConf().setAppName("dfuse").setMaster("local[*]");JavaSparkContext sc = new JavaSparkContext(conf);SQLContext sqlContext = new SQLContext(sc);JavaRDD<String> oraginRDD = sc.textFile("E:\\testdata\\students.txt");JavaRDD<Row> rowJavaRDD = oraginRDD.map(new Function<String, Row>() {@Overridepublic Row call(String s) throws Exception {String[] strs = s.split(",");return RowFactory.create(Integer.valueOf(strs[0]),strs[1],Integer.valueOf(strs[2]));}});// 定义StructType结构类型List<StructField> fields = new ArrayList<>();fields.add(DataTypes.createStructField("id",DataTypes.IntegerType, true));fields.add(DataTypes.createStructField("name",DataTypes.StringType, true));fields.add(DataTypes.createStructField("age", DataTypes.IntegerType, true));StructType type = DataTypes.createStructType(fields);Dataset ds = sqlContext.createDataFrame(rowJavaRDD,type);ds.show();}
}
Scala版本
与Java的思路相似，构造RDD<Row>，StructType进而构建DataFrame。
import org.apache.spark.sql.types.{IntegerType, StringType, StructField, StructType}
import org.apache.spark.{SparkConf, SparkContext}
import org.apache.spark.sql.{Row, SQLContext}object Programme2DF {def main(args: Array[String]): Unit = {val conf = new SparkConf().setAppName("DFUse").setMaster("local[*]")val sc = new SparkContext(conf)val sqlContext = new SQLContext(sc)import sqlContext.implicits._ // 放在sqlContext之后val rowRDD = sc.textFile("E:\\testdata\\students.txt").map(x => Row(x.split(",")(0).toInt,x.split(",")(1),x.split(",")(2).toInt))// 构造StructTypeval structType = StructType(Array(StructField("id", IntegerType, true),StructField("name", StringType, true),StructField("age", IntegerType, true)))/** 另一种写法* StructType(StructField("id",IntegerType) :: Nil)* */val df = sqlContext.createDataFrame(rowRDD, structType)df.show()}
}
最后感谢石杉老师！