Apple Numbers STOCKH Function: Historical Dates, Modes and Returns

The official STOCKH reference was retrieved October 15, 2026. The original introduction release and last functioning version are unknown from the checked first-party evidence. Current documentation is not a final supported-version endpoint or a test of every older Numbers client.

No native Numbers session was executed. Fixed examples below are independent controls; remotely supplied values are not claimed as current observations. First-party example values, when identified, remain source examples rather than newly retrieved native results.

STOCKH(symbol, close, date)

symbol is a quoted symbol or a reference to a cell containing one. The optional attribute mode is 0 for close, 1 open, 2 high, 3 low and 4 volume; the source also illustrates named attributes. Close is the documented default, but the formulas here supply the middle argument explicitly before the date.

date is a valid past date or a formula/reference providing one. Use DATE formulas when a slash-date string could be interpreted differently under another region.

The numeric modes are specific to STOCKH. For example, mode 4 is volume here, whereas STOCK uses 4 for open. Do not copy a numeric selector between the two functions without checking its meaning.

=STOCKH("AAU","volume",DATE(2015,12,15))

Apple’s historical example gives 127700 shares traded. That is a published illustration, not a fresh provider response or a claim that the symbol’s present listing status was checked.

The reference also shows an IFERROR wrapper returning text when historical data cannot be obtained. A missing observation is not a numerical zero-volume or zero-price observation merely because a report needs a placeholder.

For independent report checks, use this synthetic dataset for one imaginary instrument in a single currency:

These are not real quotations or native results. For the first row, mode targets are 0 → 52, 1 → 50, 2 → 53, 3 → 49 and 4 → 1000. This fixed mapping makes an incorrect attribute-column selection easy to detect.

The close-to-close changes are 1 and -2. Their independently calculated percentage returns are (53/52)-1, approximately 1.92307692%, and (51/53)-1, approximately -3.77358491%.

The overall close-to-close return from 52 to 51 is approximately -1.92307692%. Adding the two daily percentage returns does not reproduce that compounded two-interval return exactly; their denominators differ.

Daily high-low ranges are 4, 3 and 3.5. Total volume is 3100 shares. The mean of the three closes is 52, but it is not a volume-weighted transaction price. The dataset contains no individual trade prices from which to reconstruct such a measure.

For a fictional position of 10 shares, the three closing valuations are 520, 530 and 510. The final-minus-initial difference is -10 in the common currency.

This simple control assumes the share count stays fixed. A stock split, reverse split, dividend or trade would require additional records before the difference could be interpreted as a complete investment return.

Apple explicitly warns that splits can affect historical function output. The page does not establish a universal adjusted-price policy, so do not compare values from different services and assume they use identical adjustments.

A fixed date cell is preferable to TODAY()-2 when a report needs to keep requesting the same historical day. A relative date changes over time and can land on a date without trading or data.

The source’s error-handling example allows a text fallback when an instrument was not traded or data is unavailable. It does not promise a universal previous-business-day substitution for weekends or holidays. Do not invent one when building a time series.

Apple documents no value for a formula entered without internet and an update when connectivity returns. Some international instruments require exchange suffixes, and displayed currency depends on regional context. Exact unsupported-symbol errors, provider refresh behavior and regional parsing were not native-tested.

Keep date, symbol, attribute, denomination and retrieval context together. Those fields are necessary to distinguish a corrected historical observation, an identity mismatch and a genuine formula-reference error.