3. Vector Concepts : Data Manipulation : Granularity of DML Operations
 
Share this page                  
Granularity of DML Operations
Vector distinguishes between two granularities of DML operations:
Batch – Operations that involve a relatively small number (for example, thousands or tens of thousands) of records. Such updates are buffered in the system memory and merged with the data from disk on the fly.
INSERT, DELETE, and UPDATE are always executed as batch operations. COPY FROM and INSERT...SELECT are executed as batch operations if the destination table has a clustered index declared and is not empty.
Bulk – Operations that involve a large number of records (for example, one million or more), in which the data changes do not necessarily fit in the system memory.
CREATE TABLE AS SELECT is always executed as a bulk operation. Data load through vwload, COPY FROM and INSERT...SELECT are executed as bulk operations if the destination table has no clustered index declared or if it is empty.
For more information, see Batch and Bulk Updates.