Parcel scans after the watermark

You are a data engineer at a logistics platform. Every night an incremental pipeline picks up fresh parcel scans from the parcel_scans table and appends them to the warehouse. The position of the last successful load (the watermark) is stored in the etl_state table — in the row with source_name = 'parcel_scans' (the table may hold rows for other sources too). Your job is to build the next increment: the scans the pipeline has not seen yet, i.e. those whose time is strictly greater than the watermark. Scans with an unknown time (scan_time IS NULL) are considered not ready and must not be included. Return columns: - scan_id — scan identifier; - parcel_code — parcel code; - hub_code — sorting hub code; - scan_time — scan time (timestamp, as stored); - status — parcel status (may be NULL — output as is). Sorting: by scan_time ascending, ties broken by scan_id ascending. A scan whose time equals the watermark exactly is NOT part of the result.