Data Provenance Columns

Last updated: May 6, 2025

Note: Data provenance columns are experimental.The names of these columns or their format may be changed by Relevant without notice, which may cause the data pipeline to break. Please do not reference these columns in pipeline or report SQL without checking with your Relevant lead.

Many tables in the Relevant Data Warehouse have columns that describe the “provenance” of a row—in other words, where that row’s data comes from. Provenance information is used by data analysts when evaluating whether all appropriate data is being pulled in from source system(s).

source_system: A list of source systems that produced a given row. If a row comes from multiple source systems, the sources should be separated by a semicolon and space (e.g. “ecw; epic”).

mapping_strategy: A list of mapping strategies that produced a given row. If multiple strategies produce a row, the strategies should be separated by a semicolon and space (e.g. “visit billing codes: tobacco use cessation counseling’ value set; structured data: interested in quitting”).