Kedro - How to update a dataset in a Kedro pipeline given that a dataset cannot be both input and output of a node (only DAG)? In a Kedro project, I have a dataset in catalog.yml that I need to increment by adding a few lines each time I call my pipeline. #catalog.yml my_main_dataset: type: pandas.SQLTableDataSet credentials: postgrey_credentials save_args: if_exists: append table_name: my_dataset_name However I cannot just rely on append in my catalog parameters since I need to control that I do not insert already existing dates in my dataset to avoid duplicates. I also cannot create a node taking my...