What is ACID compliance in the context of Apache Iceberg?

ACID (Atomicity, Consistency, Isolation, Durability) compliance in Iceberg means that every write to a table — however large or complex — either fully succeeds and becomes visible as a complete, consistent new snapshot, or fully fails and leaves the table exactly as it was before, with no possibility of a reader ever seeing a partially-applied, inconsistent intermediate state.

Atomicity and durability come from Iceberg's commit mechanism: a write produces new data and metadata files, and only at the very end does the catalog's pointer atomically swap from the old metadata file to the new one — readers either see the complete old state or the complete new state, never something in between, and once committed, that new state is durably recorded in the catalog and underlying storage.

Isolation comes from snapshot immutability: because each snapshot is a fixed, unchanging view of the table, a long-running read against one snapshot is completely unaffected by writes happening concurrently on other snapshots, which is what lets Iceberg support concurrent readers and writers without the readers needing to lock the table or risk seeing inconsistent, half-written data mid-query.