Amazon Redshift announces Multidimensional Data Layouts, a new powerful table sorting mechanism that improves performance of repetitive queries. Unlike traditional sorting methods that sort data based on one or more table columns, Multidimensional Data Layouts sorts data based on the incoming query filters (for example: Sales in a specific region) , which significantly accelerates the performance of table scans. An AWS internal benchmark test shows that MDDL produced a 74% total runtime reduction compared to having no sort key, and a 40% total runtime reduction compared to the best single-column sort key.
Amazon Redshift provides automatic table optimization (ATO) capabilities which automatically optimize the design of tables by applying sort and distribution keys without the need for administrator intervention. Now, when you enable ATO or set up your table with SORTKEY AUTO, Amazon Redshift analyzes your query history and automatically selects either a single-column sort key or Multidimensional Data Layouts for your table, depending on which option is better for your workload.
Multidimensional data layouts is available as a preview in the following AWS Regions: Asia Pacific (Tokyo), US East (Ohio), US East (N. Virginia), US West (Oregon), Europe (Ireland), Europe (Stockholm). To learn more, visit the Amazon Redshift database developer guide and blog.