Guided Lesson: VForce Lakehouse Fundamentals

3. Why Apache Iceberg

Warehouse guarantees on open files

Apache Iceberg is an open table format: it lets plain data files in cloud storage behave like a real database table, so any engine can query them with strong guarantees and no vendor lock-in.

Key terms

Open table format

A published, non-proprietary way of organizing data files plus the metadata that describes them, so many different tools can read and write the same tables. The opposite of data trapped inside one vendor's warehouse.

ACID transactions

A guarantee that a set of changes either all succeed together or none do, and that readers never see a half-finished write. It is what makes a data store safe to use with many readers and writers at once.

Schema evolution

The ability to add, rename, or change columns over time without rewriting or reloading the data that is already stored.

Snapshot

A saved, read-only view of a table at a single moment. Iceberg records a new snapshot on every write.

Time-travel

Querying an earlier snapshot to see the data exactly as it looked at a past point in time, which is useful for audits and for reproducing an old report.

Catalog

The service (here, an iceberg-rest catalog) that lists the available tables and their current state so any compatible query engine can discover and read them live.