dataframe保存为csv_sparkSQL_DataFrame

进入spark安装目录：/data0/spark/spark-2.2.1-bin/bin ,执行以下shell命令：

./spark-shell --master local[2] --jars /data0/spark/spark-2.2.1-bin/jars/mysql-connector-java-5.1.15-bin.jar

val df=spark.read.option("timestampFormat", "yyyy/MM/dd HH:mm:ss ZZ").json("file:///data0/project/person.json")

保存到person2.csv目录

保存到person2.csv目录
 pDF.select("name","age").write.format("csv").option("timestampFormat", "yyyy/MM/dd HH:mm:ss ZZ").save("file:///data0/project/person2.csv")

• Schema和RDD生成DataFrame

===============对比两种方式

参见：

spark-shell & spark-sql 使用​blog.csdn.net

Spark学习之路 （十八）SparkSQL简单使用​www.cnblogs.com

Spark SQL & DataFrames​spark.apache.org

Spark SQL, Built-in Functions​spark.apache.org Spark SQL and DataFrames​spark.apache.org