Support partitions when tables are created and be able to replace a partition

[jherrmannNetfonds]

Hi everybody,

When ingesting data into a cloud data warehouse (BigQuery, Snowflake etc.) I want to partition my data (for example by execution_date), so that when I clear/rerun a DAG, only the data of the affected partition is replaced.
Currently Tables cannot automatically created with a partition. Maybe the metadata parameter of the Table Class would be a option to add something like that. When I use the append operator I should be able to replace a partition.
Are there any ideas into this direction yet?
I am not sure, what would be a good place to implement a "replacement of a partition", since the sdk has several operators which can store data into a cloud DWH, but with different option on how to store the data in the target table. For example has the load_file operator has more option to change the load behavior then a transform operator. At least as far as I understand.