For example a csv as below:
unloading_time_id,unloading_start_time,unloading_end_time,truck_id,delivery_schedule_id
1,10:00:00,13:00:00,4,1
2,12:00:00,15:00:00,2,2
3,14:00:00,17:00:00,5,3
4,16:00:00,19:00:00,1,4
5,18:00:00,21:00:00,3,5
The column unloading_start_time and unloading_end_time represents time component only. When this csv is uploaded as dataset in staging tab the type of the those columns are set as timestamp. This also the values stored in the pipeline json are stored as timestamp like 2023-09-13T10:00:00.000Z. This changes the intent of the column.
There is a workaround, to manually update the pipeline json and reload the workspace.