OceanBase Database supports direct load as an efficient method for inserting data. The core idea is to bypass the SQL parsing layer and directly allocate space and write data into the underlying data files. This feature fully leverages OceanBase’s distributed architecture and analytical processing (AP) capabilities, significantly improving the throughput and efficiency of bulk data imports while ensuring data consistency and high availability.
Two modes of direct load
Full direct load
Applicable scenarios: Used for database initialization, data migration, or whenever there is a need to quickly load large volumes of data.
AP capability advantage: OceanBase’s automatic partitioning mechanism can evenly distribute data according to the partition key, avoiding hotspots.
Incremental direct load
Applicable scenarios: Used in high-throughput data writing scenarios, such as large-scale real-time data collection, log writing, and similar cases.
AP capability advantage: After new nodes are added, OceanBase automatically performs load balancing, dynamically adjusting the distribution of data shards.