When a column is removed from a source database or dataset, the next pipeline run usually fails, making the problem easy to detect. But when a field keeps the same name and data type but its meaning c