As far as I know, you cannot reuse the direct output directory. Hadoop prevents this. if I understand correctly, you are dealing with daily logs, so I suggest you set the parent output directory, called output, and set the output directory in the script for output / daily_date.
source
share