Yahoo Finance Historical Data: Downloads, Adjustments, and Gaps
What the download contains, which assumptions it hides, and how to validate it before research.
Yahoo Finance historical data commonly provides date, open, high, low, close, adjusted close, and volume for a selected symbol and interval. Researchers must verify the interval, timezone, adjustment method, corporate actions, date boundaries, symbol history, missing sessions, and applicable terms before treating the download as a reproducible dataset.
Close and adjusted close are different tools
Raw close reflects the quoted price for that session. Adjusted close applies historical adjustments intended to make returns comparable across splits and distributions. The exact treatment matters, especially when OHLC fields and close are not adjusted in the same way by a client.
Choose the adjustment view before computing returns and save the corporate-action inputs. Mixing raw entries with adjusted exits can manufacture large false signals around splits.
Validate the download
Compare returned dates with the exchange calendar, inspect first and last rows, detect duplicates, verify the timezone, and distinguish a legitimate no-trade session from a failed request. Keep volume aligned to the same interval and source.
Current tickers can hide delisted or renamed securities. For universe backtests, use a security master and point-in-time membership rather than downloading only symbols that exist today.
- Record request parameters and retrieval time.
- Save raw and normalized copies.
- Test corporate-action dates.
- Create an expected-session coverage report.
From convenient download to durable pipeline
A production collector needs retries, rate control, schema tests, raw-object hashes, correction handling, and fallback sources. It also needs a legal decision about internal use versus downstream serving.
DataCedar separates those layers. Acquisition adapters can change while normalized tables, coverage, source lineage, authentication, and public API behavior remain under DataCedar’s control.
How DataCedar preserves the evidence
DataCedar separates acquisition from serving. Permitted source responses are retained with retrieval time and identifiers, normalized into DataCedar-owned tables, checked against expected coverage, and exposed through a stable versioned API. A collector can be replaced without changing the customer contract or making an upstream provider a runtime dependency.
Every research stream carries effective and known-at time where the distinction matters. Rights-restricted, unavailable, partial, stale, and genuinely empty states remain visible, so a backtest can fail closed and a buyer can see the product boundary before committing engineering time.
Key takeaways
- 01State raw versus adjusted values explicitly.
- 02Validate every date range against an exchange calendar.
- 03Keep symbol history and delisted securities in research universes.
- 04Do not infer serving rights from the download button.
Query public company evidence through one stable API. Free Explorer tier, no card.
Start freeView pricing