Data lineage records the path a data item takes before reaching its current table. Keep it in a data dictionary alongside attribute meaning, business rules, example data, data types, and upstream dependencies. Without understanding those business meanings, a schema can embed false assumptions. section 4.1.3
Tracing a marketing identifier
For an internet-marketing database, trace a cookie identifier from Tracking ID in an Impressions CSV supplied by a cookie provider, to TrackingID in a staging Cookies table, to CookieID in a marketing Cookies table. This makes the origin and each storage step explicit. section 4.1.3
Use lineage as part of schema design:
- Work with a business analyst or application owner to establish each attribute’s business meaning and rules before recording its path. section 4.1.3
- Record upstream dependencies and the successive names and locations of an item as it moves into its destination table. section 4.1.3
