Apache Iceberg's MERGE INTO with Merge-on-Read strategy provides significant advantages over INSERT OVERWRITE for data lake operations. MERGE INTO offers up to 66% cost savings by only modifying affected rows rather than rewriting entire partitions, reduces I/O operations, and handles partition evolution more gracefully. The approach is particularly effective for incremental loads, CDC, and slowly changing dimensions, though proper compaction maintenance is essential for long-term performance.
Table of contents
Understanding MERGE INTO vs. INSERT OVERWRITEINSERT OVERWRITEMERGE INTOPerformance benefits of MERGE INTO (MOR)Cost savings with MERGE INTO (MOR)Sort: