# Data Column-Level Lineage and Impact Analysis
Status: public
Confidence: medium (0.725) (verified)
Last verified: 2026-06-02
Generation: ai_structured


## TL;DR

Column-level lineage lets agents trace how a field in a report, feature, or RAG index was derived from upstream data.

## Core Explanation

Table-level lineage can miss the exact blast radius of a schema or transformation change. Column-level lineage tracks how individual fields flow through queries, jobs, views, and downstream assets.

Agents should use column lineage for impact analysis before renaming, deleting, masking, or changing the semantics of a field. The answer should state whether lineage is parser-derived, platform-captured, manually curated, or incomplete.

## Source-Mapped Facts

- OpenLineage documentation defines a column lineage dataset facet for describing column-level lineage. ([source](https://openlineage.io/docs/spec/facets/dataset-facets/column_lineage_facet/))
- OpenMetadata documentation describes column-level lineage as showing how columns are transformed across assets. ([source](https://docs.open-metadata.org/v1.11.x/how-to-guides/data-lineage/column))
- Databricks Unity Catalog documentation describes data lineage across tables, columns, notebooks, workflows, and dashboards. ([source](https://docs.databricks.com/aws/en/data-governance/unity-catalog/data-lineage))

## Further Reading

- [OpenLineage Column Lineage Facet](https://openlineage.io/docs/spec/facets/dataset-facets/column_lineage_facet/)
- [OpenMetadata Column-Level Lineage](https://docs.open-metadata.org/v1.11.x/how-to-guides/data-lineage/column)
- [Databricks Unity Catalog Lineage](https://docs.databricks.com/aws/en/data-governance/unity-catalog/data-lineage)