Keep the fund identity intact
An ETF and its benchmark are different records. Use the actual fund-share identifier as the option’s underlying and preserve benchmark relationships as additional context. Similar exposure is useful for comparison, but it does not establish interchangeable contracts or identical settlement.
Retain the deliverable
The OCC’s ETF specifications describe the standard contract and its exercise features. Keep each contract’s actual terms, premium scale and deliverable rather than deriving them from a fund name. Adjusted series need their verified definitions. Account-level exercise processing remains a question for the relevant provider.
Join events as evidence, not hindsight
Store verified fund events with effective dates, source references and retrieval times. A chart can annotate them without rewriting original quotes. In historical research, later event knowledge should not silently become an input that the process supposedly had earlier.
Do not sell the midpoint as a fill
A chain should distinguish bid, ask, midpoint and last trade. Keep spread, size where supplied and observation time together. Research execution assumptions belong in their own layer, with explicit costs and no-fill behavior. The linked article shows how to evaluate these choices without presenting a simulation as an observed outcome.
