Metadata Management, Catalog, and Lineage
What this domain tests
Metadata Management, Catalog, and Lineage tests how you make data discoverable and traceable. Expect questions on technical versus business metadata, lineage stitching, and catalog scope.
Core rules
- Metadata Command Center catalogs sources, scans technical metadata, and captures lineage. Business glossaries define terms, owners, and policies; the catalog links those terms to physical assets.
- Technical lineage traces columns through mappings, SQL, and jobs. Business lineage expresses the same flow in business terms. Both are needed for impact analysis.
- Lineage is stitched from multiple sources: IDMC mappings, Databricks jobs, SQL, and BI tools. Gaps appear when logic is opaque, dynamic SQL is unresolved, or a source is not scanned.
- Data Marketplace publishes curated data products with descriptions, owners, and access guidance. Catalog sync is not instantaneous; schedule scans and verify freshness.
Worked example
A finance KPI is wrong. Start at the dashboard, trace to the gold Delta table, then to the silver transformation and the IDMC mapping that loaded bronze. Column-level lineage shows that the KPI uses a renamed source column. Link the glossary term "Net Revenue" to the gold column so business users see the definition and owner.
Traps
Confusing technical with business lineage; assuming lineage appears without scanning; expecting real-time catalog updates; and cataloging tables without owners or definitions.