Daily delivery does not mean daily data

Short answer: how often a feed arrives and how often its contents change are two different facts. A dataset can be delivered every day and only gain new figures twice a year. An empty daily file is usually correct, and a monitor that treats it as a failure will cry wolf until somebody switches it off.

Notes arrive with filings, and filings are not quarterly

ChinaScope Financial Notes accompany interim and annual filings only. Expect June 30 and December 31 period-ends, not quarterly detail. (Data cut 2026-04-09.)

So the delivery is routine and the data is semi-annual. If you are diffing this feed daily and alerting on "no change", almost every alert you get will be correct behaviour rather than a fault.

One table can lag its own siblings

Inside ChinaScope Affiliate, the actual-controller table runs a period behind the rest of the dataset. It currently extends through 2024-12-31. For the most recent control picture, read the parent and controlling-shareholder table (equity_parent_company), which runs to the latest period. (Data cut 2026-05-16.)

This one is worth a check rather than an assumption, because both tables answer "who controls this company" and only one of them is current. Joining them without noticing gives you a control picture that is partly this year and partly last.

The general rule

Before you build a freshness monitor on any dataset, separate two questions. When does a file arrive, and when does the underlying disclosure change? Alert on the first. Measure staleness against the second.