Multi-statement transactions for Apache Iceberg tables in Amazon Redshift

Amazon Redshift now supports multi-statement transactions for Apache Iceberg tables, enabling multiple DML statements to commit as one atomic snapshot instead of individually.

• Wrap multiple DML statements in BEGIN...COMMIT to publish them as a single Iceberg snapshot
• Readers see the table before the batch or after it, never in an intermediate state
• Audit data quality checks inside the transaction before committing or rolling back
• All writes in a transaction must target the same Iceberg table
• Batching 25 statements reduced run time by 14.77% for INSERT and 11.29% for DELETE
• Throughput improves significantly with concurrency: 3.59× more committed statements at 32 concurrent writers
• Orphan files from failed transactions are cleaned up automatically or via table optimizers
• Available in Amazon Redshift Patch 206 and later

Multi-statement transactions make data pipelines more robust by ensuring atomicity, eliminating manual undo operations, and reducing catalog commit overhead.