In apache spark, I'm having a dataset which I'm writing to disk:
var ds = Seq(1, 2, 3).toDS();
ds.write.parquet("/tmp/01/01");
As expected, this writes records on disk in parquet format:
vaxaci$ ls -ltr /tmp/01/01/
total 24
drwxr----- 1 vaxaci hadoop 430 Oct 19 14:12 part-00000-ba320358-7ee3-46dc-b90f-3b9c94fa6503-c000.snappy.parquet
drwxr----- 1 vaxaci hadoop 430 Oct 19 14:12 part-00001-ba320358-7ee3-46dc-b90f-3b9c94fa6503-c000.snappy.parquet
drwxr----- 1 vaxaci hadoop 430 Oct 19 14:12 part-00002-ba320358-7ee3-46dc-b90f-3b9c94fa6503-c000.snappy.parquet
drwxr----- 1 vaxaci hadoop 0 Oct 19 14:12 _SUCCESS
But the problem is with the UNIX group with which it's writing. Currently it is writing with group name hadoop.
Is there any way through which I can programatically (in java preferably) modify this group to something else, say friends?