Merge four CSVs locally, robustly and efficiently

You are building a local data pipeline that merges four CSV files into two CSV reports. The laptop has 16 GiB of RAM, and no cloud or network resources may be used. purchases.csv can contain up to 15 million rows, so the implementation must stream or spill to disk instead of loading the entire file into memory. The source files are: products.csv: product_id, category, product_name purchases.csv: purchase_id, product_id, price, stars ads_events.csv: event_date, user_id,…