Spark SQL——JSON数据源

• Spark SQL可以自动推断JSON文件的元数据，并且加载其数据，创建一个DataFrame。可以使用SQLContext.read.json()方法，针对一个元素类型为String的RDD，或者是一个JSON文件。

• 但是要注意的是，这里使用的JSON文件与传统意义上的JSON文件是不一样的。每行都必须，也只能包含一个，单独的，自包含的，有效的JSON对象。不能让一个JSON对象分散在多行。否则会报错。

综合性复杂案例：查询成绩为80分以上的学生的基本信息与成绩信息

基于java

package cn.spark.study.sql;

import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaPairRDD;
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.PairFunction;
import org.apache.spark.sql.DataFrame;
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 scala.Tuple2;
import java.util.ArrayList;
import java.util.List;
/**
 * JSON数据源
 */
public class JsonDataSource_9 {
    public static void main(String[] args) {
        SparkConf conf = new SparkConf()
                .setMaster("local")
                .setAppName("JsonDataSource_9");
        JavaSparkContext sc = new JavaSparkContext(conf);
        SQLContext sqlContext = new SQLContext(sc);

        //一：对json文件，创建DataFrame
        DataFrame studentScoreDF = sqlContext.read().json("E:\\sparktext\\students-score.json");
        //查询学生成绩大于80的人
        studentScoreDF.registerTempTable("student_score");
        DataFrame goodsStudentScoreDF = sqlContext.sql("select name,score from student_score where score >= 80");
        List<String> goodsStudentNames = goodsStudentScoreDF.javaRDD().map(new Function<Row, String>() {
            @Override
            public String call(Row row) throws Exception {
                return row.getString(0);
            }
        }).collect();


        //二：针对JavaRDD<String>创建DataFram
        //针对包含json串的JavaRDD，创建DataFrame
        List<String> studentInfosJson = new ArrayList<String>();
        //{"name":"Leo", "score":85}
        studentInfosJson.add("{\"name\":\"Leo\", \"age\":18}");
        studentInfosJson.add("{\"name\":\"Marry\", \"age\":17}");
        studentInfosJson.add("{\"name\":\"Jack\", \"age\":19}");
        JavaRDD<String> StudentInfoJsonRDD = sc.parallelize(studentInfosJson);
        DataFrame studentInfoDF = sqlContext.read().json(StudentInfoJsonRDD);


        //针对学生基本信息DataFrame注册临时表，然后查询分数大于80分的学生的基本信息
        studentInfoDF.registerTempTable("student_info");
        //select name,age from student_info where name in ('leo' ,'marry')
        String sql = "select name,age from student_info  where name in (";
        for (int i = 0;i<goodsStudentNames.size();i++){
            sql += "'" + goodsStudentNames.get(i) +"'";
            if(i < goodsStudentNames.size() -1){
                sql += ",";
            }
        }
        sql += ")";
        DataFrame goodStudentInfoDF = sqlContext.sql(sql);

        /**
         * +-----+-----+-----+---+
         * | name|score| name|age|
         * +-----+-----+-----+---+
         * |  Leo|   85|  Leo| 18|
         * |  Leo|   85|Marry| 17|
         * |Marry|   99|  Leo| 18|
         * |Marry|   99|Marry| 17|
         * +-----+-----+-----+---+
         */
        //DataFrame join = goodsStudentScoreDF.join(goodStudentInfoDF);
        //然后将两份数据的DataFrame，转换成JavaPairRDD，执行join transformation
        //将DataFrame转换成JavaRDD，在map为JavaPairRDD，然后进行join
        //(name,score)
         JavaPairRDD<String, Integer> goodsStudentScoreTuple = goodsStudentScoreDF.javaRDD().mapToPair(new PairFunction<Row, String, Integer>() {
            @Override
            public Tuple2<String, Integer> call(Row row) throws Exception {
                return new Tuple2<String, Integer>(row.getString(0),
                        Integer.valueOf(String.valueOf(row.getLong(1))));
            }
        });

        //(name,age)
        final JavaPairRDD<String, Integer> goodStudentInfoTuple = goodStudentInfoDF.javaRDD().mapToPair(new PairFunction<Row, String, Integer>() {
            @Override
            public Tuple2<String, Integer> call(Row row) throws Exception {
                return new Tuple2<String, Integer>(row.getString(0),
                        Integer.valueOf(String.valueOf(row.getLong(1))));
            }
        });

        JavaPairRDD<String, Tuple2<Integer, Integer>> join = goodsStudentScoreTuple.join(goodStudentInfoTuple);

        //然后将封装在RDD中是好学生的全部信息，转换成一个JavaRDD<Row>的格式
        JavaRDD<Row> goodStudentRowRDD = join.map(new Function<Tuple2<String, Tuple2<Integer, Integer>>, Row>() {
            @Override
            public Row call(Tuple2<String, Tuple2<Integer, Integer>> tuple2) throws Exception {
                return RowFactory.create(tuple2._1, tuple2._2._1, tuple2._2._2);
            }
        });

        //创建一份元数据，将JavaRDD<Row>转换成DataFrame

        List<StructField> studentFields = new ArrayList<StructField>();
        studentFields.add(DataTypes.createStructField("name", DataTypes.StringType, true));
        studentFields.add(DataTypes.createStructField("score", DataTypes.IntegerType, true));
        studentFields.add(DataTypes.createStructField("age", DataTypes.IntegerType, true));

        StructType structType = DataTypes.createStructType(studentFields);

        //将JavaRDD转换成DataFrame
        final DataFrame dataFrame = sqlContext.createDataFrame(goodStudentRowRDD, structType);
        dataFrame.write().format("json").save("E:\\sparktext\\good_student_score_java");




//        join.foreach(new VoidFunction<Tuple2<String, Tuple2<Integer, Integer>>>() {
//            @Override
//            public void call(Tuple2<String, Tuple2<Integer, Integer>> tuple) throws Exception {
//                System.out.println("name:" + tuple._1);
//                System.out.println("score:" + tuple._2._1+":"+ "age:" + tuple._2._2);
//            }
//        });
    }
}

基于scala

package cn.spark.study.sql

import org.apache.spark.rdd.RDD
import org.apache.spark.sql.types.{IntegerType, StringType, StructField, StructType}
import org.apache.spark.{SparkConf, SparkContext}
import org.apache.spark.sql.{DataFrame, Row, SQLContext}

/**
  * 两种都去读取json的方式创建DataFrame
  * 1.直接读取json文件
  * 2.加载RDD但是RDD中的元素是满足json格式的String类型
  */
object JsonDataSource_9 {

  def main(args: Array[String]): Unit = {
    val conf = new SparkConf()
      .setMaster("local")
      .setAppName("JsonDataSource_9")
    val sc = new SparkContext(conf)
    val sqlContext = new SQLContext(sc)

    //创建学生成绩DataFrame
    val studentScoreDF: DataFrame = sqlContext.read.json("E:\\sparktext\\students-score.json")
    //查询分数大于80的学生
    val goodStudentScoreDF :DataFrame= studentScoreDF.filter(studentScoreDF.col("score") >= 80)

//    goodStudentScoreDF.foreach(row => {
//      println(row.getString(0)+","+row.getLong(1))
//    })

    val goodStudentNames: Array[String] = goodStudentScoreDF.map(row => row.getString(0)).collect()
//
//    //创建学生基本信息数据
    val studentInfosJSONs = Array("{\"name\":\"Leo\", \"age\":18}",
      "{\"name\":\"Marry\", \"age\":17}",
      "{\"name\":\"Jack\", \"age\":19}")

    val studentInfoJsonRDD: RDD[String] = sc.parallelize(studentInfosJSONs,1)
    val studentInfoDF: DataFrame = sqlContext.read.json(studentInfoJsonRDD)

//
//    //查询出学生成绩大于80分的基本信息
    studentInfoDF.registerTempTable("student_info")
    var sql = "select name,age from student_info where name in ("
    for( i <- 0 until goodStudentNames.length){
      sql += "'" +goodStudentNames(i)+"'"
      if(i < goodStudentNames.length-1){
        sql += ","
      }
    }
    sql += ")"
    val goodStudentInfoDF: DataFrame = sqlContext.sql(sql)

//    goodStudentInfoDF.foreach(row => {
//      println(row.getString(0)+","+row.getLong(1))
//    })
//
    val goodStudentRDD: RDD[(String, (Long, Long))] = goodStudentScoreDF.rdd.map(row => (row.getAs[String]("name"), row.getAs[Long]("score")))
      .join(goodStudentInfoDF.rdd.map(row => (row.getAs[String]("name"), row.getAs[Long]("age"))))
    val goodStudentRowRDD: RDD[Row] = goodStudentRDD.map(info => {
      //注意数据类型与67和68行的元数据类型匹配
      Row(info._1, info._2._1.toInt, info._2._2.toInt)
    })

    val structType = StructType(Array(
      StructField("name",StringType,true),
      StructField("score",IntegerType,true),
      StructField("age",IntegerType,true)))

    val goodStudentsDF: DataFrame = sqlContext.createDataFrame(goodStudentRowRDD,structType)
    goodStudentsDF.write.format("json").save("E:\\sparktext\\good_student_score_scala")
  }
}