11.6 hdfs读取json数据，转成DataFrame保存到hdfs

读取json或者parquet文件创建一个DataFrame

DataFrame存储到某一个路径下，默认存储格式是parquet

SaveMode.Overwrite：重写

SparkConf conf = new SparkConf()

.setAppName("SaveModeTest")

.setMaster("local");

JavaSparkContext sc = new JavaSparkContext(conf);

SQLContext sqlContext = new SQLContext(sc);

//去hdfs读取文件，返回DataFrame格式

DataFrame peopleDF = sqlContext.read().json("hdfs://hadoop1:9000/input/people.json");

//存储保存到hdfs上，以parquet格式

peopleDF.write().mode(SaveMode.Overwrite).save("hdfs://hadoop1:9000/output/namesAndFavColors_scala");

//验证是否parquet，就是以该格式读取显示

sqlContext.read().format("parquet").load("hdfs://hadoop1:9000/output/namesAndFavColors_scala").show();

// sqlContext.read().parquet("hdfs://hadoop1:9000/output/namesAndFavColors_scala").show();