java

 1 public class DynamicDemo {
 2     private static SparkConf conf = new SparkConf().setAppName("dynamicdemo").setMaster("local");
 3     private static JavaSparkContext jsc = new JavaSparkContext(conf);
 4     private static SparkSession session = new SparkSession(jsc.sc());
 5 
 6     public
 
 static void main(String[] args) {
 7 
 8         // 創建rdd
 9         JavaRDD<String> rdd = jsc.textFile("./src/main/java/cn/tele/spark_sql/rdd2dataset/students.txt");
10 
11         // 創建Row的rdd
12         JavaRDD<Row> rowRdd = rdd.map(new Function<String, Row>() {
13 
14             private static
 
 final long serialVersionUID = 1L;
15 
16             @Override
17             public Row call(String v1) throws Exception {
18                 String[] fields = v1.split(",");
19                 return RowFactory.create(Integer.valueOf(fields[0]), fields[1], Integer.valueOf(fields[2]));
20             }
 
21         });
22 
23         // 創建schema
24         StructType schema = DataTypes
25                 .createStructType(Arrays.asList(DataTypes.createStructField("id", DataTypes.IntegerType, false),
26                         DataTypes.createStructField("name", DataTypes.StringType, false),
27                         DataTypes.createStructField("age", DataTypes.IntegerType, false)));
28 
29         // 轉換
30         Dataset<Row> dataset = session.createDataFrame(rowRdd, schema);
31 
32         dataset.createOrReplaceTempView("students");
33 
34         Dataset<Row> result = session.sql("select * from students where age<=18");
35         result.show();
36 
37         jsc.close();
38     }
39 }

scala

 1 object DynamicDemo {
 2   def main(args: Array[String]): Unit = {
 3     val conf = new SparkConf().setAppName("reflectdemo").setMaster("local")
 4 
 5     val sc = new SparkContext(conf)
 6 
 7     val sqlContext = new SQLContext(sc)
 8 
 9     //創建rdd
10     val rdd = sc.textFile("./src/main/scala/cn/tele/spark_sql/rdd2dataframe/students.txt", 8)
11 
12     val rowRdd = rdd.map(lines => {
13       val arr = lines.split(",");
14       Row(arr(0).trim().toInt, arr(1), arr(2).trim().toInt)
15     })
16 
17     val schema = DataTypes.createStructType(Array(
18       /*    DataTypes.createStructField("id",DataTypes.IntegerType,false),
19           DataTypes.createStructField("name",DataTypes.StringType,false),
20           DataTypes.createStructField("age",DataTypes.IntegerType,false)*/
21       StructField("id", DataTypes.IntegerType, false),
22       StructField("name", DataTypes.StringType, false),
23       StructField("age", DataTypes.IntegerType, false)))
24 
25     //轉換
26     val dataframe = sqlContext.createDataFrame(rowRdd, schema)
27 
28     dataframe.createOrReplaceTempView("students")
29 
30     val result = sqlContext.sql("select * from students where age<=18")
31     result.show()
32   }
33 }

sparksql 動態設置schema將rdd轉換成dataset/dataframe