help='Max size of in-memory data for each partition in a single hive dataset. Smaller partitions cause more files to be written. Can pass the byte size as an integer or as a human-readable byte string. For example, 1024 and 1KiB are equivalent.')
parser.add_argument('--no-clobber',
type=bool,
default=False,
action='store_true',
help="Flag to set whether contents of a hive cell will be overwritten. If True, the pipeline will exit if any parquet files are found in the cell directory. No processing will occur in that case. If False (default), any files existing in the cell directory will be removed prior to data writing.")