Which operation is most appropriate to rewrite small Delta files into larger, more scan-efficient files to reduce the number of files and improve performance?

Which operation is most appropriate to rewrite small Delta files into larger, more scan-efficient files to reduce the number of files and improve performance?

• A. VACUUM
• B. ANALYZE
• C. OPTIMIZE
• D. REBUILD

Answer: (C)

Fewer, larger files improve scanning because there’s less metadata to manage and fewer file-open operations per query. The operation that achieves this on Delta tables is OPTIMIZE, which rewrites many small Delta files into larger ones. This coalescing reduces the number of files that must be read and can greatly speed up scans, especially after ingest pipelines that produce lots of tiny files. You can also pair OPTIMIZE with Z-Ordering to cluster related data for even faster range scans. VACUUM removes old or unused files but doesn’t reorganize current data, ANALYZE collects statistics for query planning but doesn’t change file sizes, and REBUILD isn’t the standard tool for this file-coalescing goal. So OPTIMIZE is the best choice to rewrite small Delta files into larger, more scan-efficient ones.

Fewer, larger files improve scanning because there’s less metadata to manage and fewer file-open operations per query. The operation that achieves this on Delta tables is OPTIMIZE, which rewrites many small Delta files into larger ones. This coalescing reduces the number of files that must be read and can greatly speed up scans, especially after ingest pipelines that produce lots of tiny files. You can also pair OPTIMIZE with Z-Ordering to cluster related data for even faster range scans. VACUUM removes old or unused files but doesn’t reorganize current data, ANALYZE collects statistics for query planning but doesn’t change file sizes, and REBUILD isn’t the standard tool for this file-coalescing goal. So OPTIMIZE is the best choice to rewrite small Delta files into larger, more scan-efficient ones.