[ARROW-11928] [C++][Compute] Add ExecNode hierarchy

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 5.0.0
• Component/s: C++
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/27765
• Assignee: Ben Kietzman
• Reporter: Ben Kietzman
• Created: 10/Mar/21 22:33
• Updated: 11/Jan/23 08:23
• Resolved: 11/May/21 16:51

Description

Per discussion on https://docs.google.com/document/d/1AyTdLU-RxA-Gsb9EsYnrQrmqPMOYMfPlWwxRi1Is1tQ

Add an ExecNode interface with which a streaming execution graph can be constructed. Initial concrete classes will include:

• ScanNode, which wraps a dataset and is a pure emitter of batches (initially, this will only wrap memory sized datasets such as tables. See ARROW-11930)
• FilterNode, which evaluates an expression on inputs and based on the result removes rows from batches (eventually, this may defer materialization of the selection to other kernels. See ARROW-5005 ARROW-10474)
• ProjectNode, which evaluates expressions on inputs producing new columns.
• GroupedAggregateNode, which computes aggregations grouped on one or more keys.

Issue Links

• is a child of: ARROW-8894 [C++] C++ array kernels framework and execution buildout (umbrella issue)
• is a child of: ARROW-12633 [C++] Query engine umbrella issue
• supercedes: ARROW-7878 [C++] Implement LogicalPlan and LogicalPlanBuilder
• links to: GitHub Pull Request #9742
• links to: GitHub Pull Request #10204