- The cardinality and size of the duplicated table.
If there is a geospatial index defined on only one of the tables involved in the join operation, the Optimizer will duplicate the non-indexed table.
If there is a geospatial index on both tables, the optimizer will calculate the costs of duplicating each table and using the index of the other table, and choose the most cost-effective combination.
- The cardinality and size of the row ID spool.
The number of rows populating the row ID spool depends on the selectivity of the join.
The row size of a row lying within the row ID spool depends on the number of columns belonging to the duplicated table, which were referenced in the query.
For more information on how the Teradata Optimizer chooses execution paths for queries, see Teradata Vantage™ - SQL Request and Transaction Processing, B035-1142.