What is the recommended approach to versioning datasets in Fabric?

• A. Maintain versioning solely with time travel; do not tag versions.
• B. Use version tags and rely on the catalog alone with no time travel.
• C. Maintain dataset versioning with version tags, track changes in the catalog, and use time travel to access historical versions; coordinate with pipelines.
• D. Keep separate copies for each version without catalog integration.

Answer: (C)

Versioning datasets in Fabric is best achieved by combining explicit version tags, tracking changes in the data catalog, and using time travel to access historical states, all coordinated with pipelines. Version tags give you a stable, human-friendly reference to a specific state of a dataset, so downstream processes and analyses can reliably target a known version. Time travel lets you retrieve and compare past versions when you need to reproduce results, audit changes, or troubleshoot issues. The catalog keeps metadata and lineage centralized, helping you discover, govern, and coordinate dataset usage across projects and pipelines. Coordinating this with pipelines ensures that every job consumes the intended version, maintaining reproducibility and reducing drift. Relying only on time travel misses a stable reference to cite in code and schedules; using the catalog alone loses the ability to retrieve exact historical states; keeping separate copies for each version creates governance, storage, and maintenance challenges without centralized metadata and lineage.