Optimizer Enhancements in Release 16.20 Update 1 - Teradata Database - Teradata Vantage NewSQL Engine

Teradata Vantageā„¢ NewSQL Engine Release Summary

Teradata Database
Teradata Vantage NewSQL Engine
Release Number
March 2019
English (United States)
Last Update
Product Category
Teradata Vantage

These are the significant enhancements to the Teradata Database Optimizer in this release.

AJI can be Used for COUNT(*), Even Without GROUP BY Clause

Aggregate join indexes can now be used for queries that contain COUNT expressions but no GROUP BY clause.

Compressed Join Index Can Now be Used in Plans Involving Outer Joins

Compressed join indexes can now be considered for use in plans for queries involving outer joins to improve performance of these queries.

Dynamic Parameterization of Literals

When eligible nonparameterized requests differ only in the eligible literal values used in the predicates of WHERE and ON clauses, the Parser can automatically parameterize these literals. This allows the Parser to reuse cached generic execution plans for such queries, and speed up processing.

To enable this feature, you must change the value of the DBS Control field, EnableDynamicParameterization. For more information about DBS Control, see Teradata Vantageā„¢ - Database Utilities , B035-1102 .

Support for Direct PI-PI Joins between LATIN and UNICODE Columns

The Optimizer can now directly join PI columns, even if one column is defined to use the LATIN character set and the other is defined to use UNICODE.

Use Unenforced Referential Integrity Constraints to Eliminate Outer Joins

The Optimizer can now consider unenforced referential integrity constraints to do join elimination on outer joins, in addition to inner joins. Unenforced referential integrity constraints (sometimes called soft referential integrity) are those defined with a REFERENCES clause that includes WITH NO CHECK OPTION. Because these constraints are not enforced by Teradata Database, it is the responsibility of the user to make sure the constraints are honored.