Compaction in Cloudera Data Warehouse
You understand the importance of compaction and the consequences of neglecting to perform compaction. Compaction keeps your Data Warehouse healthy.
Over time tables belonging to a workload become fragmented due to operations performed on them by your workload users. These small, obsolete files might lead to performance degradation and query latency problems. Compaction plays a major role in improving response time to workload queries by reducing the number of underlying files for a table and eliminating the obsolete ones. Compaction runs periodically in the background to maintain the optimal state.
Running periodic compaction is a best practice for the performance for ACID transactions. ACID inserts and deletes generate the problematic files that you might need to monitor and manage. In Cloudera Data Warehouse (CDW), compaction is always performed by a Hive Virtual Warehouse.