What if the schema changed upstream in Hive. Sync is supposed to be idempotent and intended to run multiple times. _Originally posted by @zpappa in https://github.com/databrickslabs/ucx/pull/295#discussion_r1339133804_