All indexes (table row ids sorted by one or more fields) created in calculations are kept for possible reuse in future calculations. Indexes (and joins, if joins are used) are the most costly data to recreate, so it is important to keep them for optimal performance.

Default for server-side DataEngine. All indexes and joins created in calculations are kept for possible reuse in future calculations. They are the most costly data to recreate, so it is important to keep them for optimal performance.

All files created in calculations are deleted when no longer needed. If they are requested again, they are recreated. Use this strategy only for calculations peformed once or rarely repeated, or if saving disk space is more important than performance.

Default for FlexPivot control. All files created in calculations are kept on disk, so they can be used in future calculations without recreating them. This strategy is appropriate for FlexPivot control, but usually not for server-side DataEngine serving multiple clients.