Python Actor pattern: paid rows, free diagnostics, and replay-safe imports

[ichi08]

Disclosure: I maintain jpmarketdata’s paid Yahoo! Auctions Japan Actor. This example shows a small Python Actor data-contract pattern.

The Actor uses the Apify Python SDK to deliver billable auction rows, while keeping empty-search and charge-limit notices as free search_summary rows. A successful run can still have partial coverage, so a downstream importer should inspect those notices before treating the dataset as a complete snapshot.

The existing SQLite example imports an already completed run; it does not start another billable run. It retains auction IDs and run identity, making a replay safe while preserving distinct observed price/status variants. “First observed” is not necessarily “newly listed,” and disappearance is not evidence of a sale.

The Actor page includes the output contract and runnable importer: https://apify.com/jpmarketdata/yahoo-auction-listings

For SDK users building mixed paid-data/free-diagnostic datasets: how do you make completeness and replay behavior explicit to downstream consumers?

Actor event charges are $0.0015 per delivered auction row, including rows seen on previous runs; Apify platform resource charges may also apply. The output is a bounded search snapshot.