current export flow supports {bucket}/{date}/{table}/{file_name}.parquet in Redshift or BigQuery it's impossible to use external partitioned tables in current format, there needs to rename/move files to a proper folder structure. if it would be {bucket}/{table}/dt={date}/{file_name}.parquet ... gs://.../person/dt=2025-02-05/part-00... gs://.../person/dt=2025-03-05/part-00... gs://.../education/dt=2025-02-05/part-00... gs://.../education/dt=2025-03-05/part-00... .... there will be possibility to use a direct files as native Hive partitions.