Which operation should you schedule to combine small Delta Lake files into larger ones?

Which operation should you schedule to combine small Delta Lake files into larger ones?

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

Answer: (B)

Reducing the number of small data files improves read performance by lowering the per-file I/O and metadata overhead. The best way to do this in Delta Lake is to run the OPTIMIZE command, which rewrites the table’s data into larger Parquet files and updates the transaction log so queries see the new layout. If you want even faster access for certain filters, you can add ZORDER BY to co-locate related data in the same files. This maintenance task directly addresses the issue of many tiny files. VACUUM cleans up files no longer referenced by the table, ANALYZE gathers statistics for the optimizer, and REORG isn’t used for this file-size consolidation.

Reducing the number of small data files improves read performance by lowering the per-file I/O and metadata overhead. The best way to do this in Delta Lake is to run the OPTIMIZE command, which rewrites the table’s data into larger Parquet files and updates the transaction log so queries see the new layout. If you want even faster access for certain filters, you can add ZORDER BY to co-locate related data in the same files. This maintenance task directly addresses the issue of many tiny files. VACUUM cleans up files no longer referenced by the table, ANALYZE gathers statistics for the optimizer, and REORG isn’t used for this file-size consolidation.