two things stood out to me in the code. listings without a price will break the price formatting, even though the schema allows null, so it's safer to format only when a price exists. the script also doesn't check whether the run succeeded before reading the dataset so it can report success and then read an empty or partial result. the performance numbers would be easier to trust with sample sizes, test conditions and dates. i'd also like to see a quick check on an unpriced listing, a zero-result query and one search against the live OLX page, it would really strenghen the tutorial
Great catch, @OnlineProxy! You're completely right on both points:
price: null), evaluatingad.get('price'):,raises aTypeErrorin Python. A defensive formatter likef"R$ {price:,.0f}".replace(",", ".") if price is not None else "Sob consulta"avoids any crash and respects Brazilian currency formatting.if run.get("status") != "SUCCEEDED":before touching the dataset prevents reading empty/partial results if the Actor timed out or failed.I've updated the script in the guide to incorporate defensive null checks across price, city/state, and run status. Appreciate the sharp peer review!