@sudhir reddy Try with csv writer as text expects only one column. df.write.mode("overwrite").csv("/user/root/JOUT") (or) Convert the df to rdd then use saveAsTextFile to write the json to hdfs df.rdd.saveAsTextFile("/user/root/JOUT") (or) using .format("json") and save the file to hdfs df.write.format("json").mode("append").save("/user/root/JOUT") Some useful links regarding similar errors link1,link2,link3 ... View more

@sudhir reddy Can you try switching to "hive" user and then run the hive command? # su - hive # hive . Or please make sure to have the "/user/root" directory created in your HDFS. # su - hdfs -c "hdfs dfs -mkdir /user/root" # su - hdfs -c "hdfs dfs -chown root:hadoop /user/root" . ... View more