Getting started with Apache Iceberg write support in Amazon Redshift – Part 3

This article demonstrates schema evolution and partition management for Apache Iceberg tables in Amazon Redshift, including metadata-only ALTER operations and cross-engine access through Lake Formation resource links.

• ALTER TABLE operations (add/drop/rename columns, widen types) execute as metadata-only changes without rewriting data
• Partition evolution allows changing partition specs (add/drop/replace fields) while existing data remains in original layout
• Supported partition transforms include year, month, day, hour, bucket, truncate, and identity functions
• Lake Formation resource links enable centralized governance and cross-engine access via external schemas
• Query S3 Tables using two-part notation through external schemas instead of three-part notation
• Multi-level partition specs can be built incrementally by adding partition fields one at a time
• Best practices include testing in non-production, using REPLACE PARTITION FIELD for atomic operations, and monitoring with SHOW TABLE

Schema evolution in Iceberg enables safe production table modifications without pipeline rebuilds or data rewrites, with automatic propagation across all query engines.