6. Stage 4--Optimize the Database Schema : Performance Considerations : Ordered Data
Share this page                  
Ordered Data
VectorH makes use of storage tracking structures to record details of the range of values that may be held in a particular block. These structures are used at runtime to effect block elimination based on query predicates. This approach improves query performance generally and is particularly effective when dealing with ordered data.
Most systems have such an ordering of data—for example carrying an “inserted date” or “effective date” field that matches the date when the data was created. Mimicking such an ordering in the test system by loading the data in the appropriate order will give a better indication of query performance in the real system.
The Actian GitHub site has a tool to help determine whether the data is being stored on-disk in sorted order for a given column.
The simplest way to guarantee that data is stored in sorted order on disk is to use an index.