Lookup Table

A lookup table is a mapping that, at report time, enriches a raw value with more attributes — for example turning a product id into its category, price and brand.

A lookup table solves a practical problem: you do not want to send every attribute of an entity inside each event, because it bloats the event and, if the price changes, the old data stays wrong. Instead you send only an id and keep the attributes separately in a table that is joined to the event at report time. Its big advantage is that fixing the table updates the interpretation of all past and future data at once, with no change to instrumentation. But that is also a trap: because the mapping is applied at report time, changing the table changes the meaning of the past too, and any id missing from the table leaves those events uncategorized. Its difference from a computed property is that a lookup table is a fixed, data-driven mapping rather than conditional logic.