[ARROW-10051] [C++][Compute] Make aggregate kernel merge state mutable

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.0.0
• Component/s: C++
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/26072
• Assignee: Yibo Cai
• Reporter: Yibo Cai
• Created: 21/Sep/20 05:09
• Updated: 11/Jan/23 08:10
• Resolved: 24/Sep/20 09:39

Description

Aggregate kernel consumes each batch and output a state, which is then merged with other batches. Currently, state parameter is defined as const KernelState& in merge interface. It may cause unnecessary data copying for kernels with non-trivial states.
E.g., mode kernel maintains a value:count map in its state structure, it's better to move the map than copy it (for merge operation of the first batch).

Aggregate kernel code template is at [1]. batch_state is an auto variable(line 873), modifying it in kernel->merge() is okay (line 888).

[1] https://github.com/apache/arrow/blob/a17717fa528bb7d495cfba5c6b5fb67866934a63/cpp/src/arrow/compute/exec.cc#L871-L891

Issue Links

• links to: GitHub Pull Request #8232