spark部分json從檔案讀取資料【java純程式碼】
package com.bjsxt;
import org.apache.hadoop.hdfs.server.namenode.dfsclusterhealth_jsp;
import org.apache.spark.SparkConf;
import org.apache.spark.SparkContext;
import org.apache.spark.sql.DataFrame;
import org.apache.spark.sql.SQLContext;
public class JavaRdd {
public static <Row> void main(String[] args) {
SparkConf conf=new SparkConf().setAppName("test").setMaster("local");
SparkContext sc=new SparkContext(conf);
SQLContext SQLContext=new SQLContext(sc);
//DataFrame df = SQLContext.read().format("json").load("./json");
DataFrame df = SQLContext.read().json("./json");
df.show(100);
System.out.println("+++++++++++++++++++++++++++");
df.printSchema();
System.out.println("+++++++++++++++++++++++++++");
//JavaRDD<Row> javaRDD = df.javaRDD();
df.select("name").show();
System.out.println("___________________");
df.select(df.col("name"),df.col("age").plus(10).alias("asa")).show();
System.out.println("*********************************");
df.select(df.col("name"),df.col("age")).where(df.col("age").gt(19)).show();
System.out.println("????????????????????????//");
df.groupBy(df.col("age")).count().show();
System.out.println("PPPPPPPPPPPPPPPPPPPPP");
df.registerTempTable("jtb");
DataFrame sql = SQLContext.sql("select name,age,count(*) as gg from jtb group by age");
sql.show();
sc.stop();
}
}