A CSV cleanup rule can silently drop important rows. I built a synthetic six-row Python demo that keeps the original file, writes a clean output, and puts every rejected row in an exceptions report with a reason. Two rows are kept; four are explained: duplicate key, invalid...
