[ARROW-6690] [Rust] [DataFusion] HashAggregate without GROUP BY should use SIMD

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.16.0
• Component/s: Rust, Rust - DataFusion
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/23037
• Assignee: Andy Grove
• Reporter: Andy Grove
• Created: 25/Sep/19 13:58
• Updated: 11/Jan/23 07:48
• Resolved: 12/Oct/19 16:35

Description

Currently the implementation of HashAggregate in the new physical plan uses the same logic regardless of whether a grouping expression is used.

For the case where there is no grouping expression, such as "SELECT SUM(a) FROM b" we can use the compute kernels to perform an aggregate operation on each batch rather than iterating over each row and accumulating individual values.

This optimization already exists in the original implementation of aggregate queries direct from the logical plan.

Issue Links

• links to: GitHub Pull Request #5606