I am currently working on the Snowplow data modeling pipeline using SQL runner. I met a problem. Each time we add or update a self-describing schema, then the data modeling pipeline would fail since the target table doesn’t match the source table anymore. For now, I will delete the whole derived and scratch dataset and then create new ones manually. But I found the operation is quite frequent and dull. So I am wondering whether you have any good idea about this problem. Thanks!