Db2 for z/OS and its ecosystem

 View Only

How do organizing keys work in Db2 Analytics Accelerator and how they differ in V5 and V7

By Ute Baumbach posted Fri August 07, 2020 05:46 AM

  
An organizing key can have an impact on the query response time. When choosing an organizing key, you specify columns of an accelerator-shadow table or accelerator-archive table for the clustering of the table records. This creates grouped segments or blocks of rows with equal or nearby values in the columns selected as organizing keys. If an incoming SQL query references one of the organizing key columns in a range or an equality predicate, the query can run much faster because entire blocks can be skipped during the table scan if the underlying database system can determine that none of the records in a block satisfy the predicate.

The concept of an organizing key has existed in Db2 Analytics Accelerator for many years. It exists in the previous accelerator generation (which is Db2 Analytics Accelerator V5 and earlier versions) and in the current accelerator generation (which is Db2 Analytics Accelerator V7).
However, the feature has been implemented differently in either version. The implementation differences make it necessary to also handle organizing keys in a different way, depending on your product version.

The following white paper describes the implementation differences in the previous and current accelerator generation and describes best practices for defining, changing and maintaining organizing keys on accelerator-shadow tables and accelerator-archive tables: How do organizing keys work in Db2 Analytics Accelerator and how they differ in V5 and V7

The white paper is useful for Accelerator users who are already familiar with organizing keys in product version 5, and want to continue to use them in version 7. It is also useful for Accelerator users who want to start using organizing keys in either product.


#Db2forz/OS
#IDAA
0 comments
11 views

Permalink