强调它与方法一的区别:当DataFrame的数据结构不能够被提前定义。例如:(1)记录结构已经被编码成字符串 (2) 结构在文本文件中,可能需要为不同场景分别设计属性等
以上情况出现适用于以下方法。
1.people.txt:
soyo8, 35
小周, 30
小华, 19
soyo,88
/*** Created by soyo on 17-10-10.* 使用编程方式定义RDD模式*/import org.apache.spark.sql.types._import org.apache.spark.sql.{Row, SparkSession}object RDD_To_DataFrame2 {def main(args: Array[String]): Unit = {val spark=SparkSession.builder().getOrCreate()val peopleRDD=spark.sparkContext.textFile("file:///home/soyo/桌面/spark编程测试数据/people.txt")val schema_S="name age"val fields=schema_S.split(" ").map(x=>StructField(x,StringType,nullable = true))//生成模式val schema=StructType(fields)val rowRDD=peopleRDD.map(_.split(",")).map(x=>Row(x(0),x(1).trim))val peopleDF=spark.createDataFrame(rowRDD,schema)peopleDF.createOrReplaceTempView("people2")val results=spark.sql("select * from people2")results.show()results.groupBy("age").count().show()}}
结果:
+-----+---+
| name|age|
+-----+---+
|soyo8| 35|
| 小周| 30|
| 小华| 19|
| soyo| 88|
+-----+---+
+---+-----+
|age|count|
+---+-----+
| 30| 1|
| 35| 1|
| 19| 1|
| 88| 1|
+---+-----+