[ARROW-8500] [C++] Use selection vectors in Filter implementation for record batches, tables

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.0.0
• Component/s: C++
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/24672
• Assignee: Wes McKinney
• Reporter: Wes McKinney
• Created: 17/Apr/20 16:25
• Updated: 11/Jan/23 08:00
• Resolved: 18/Jun/20 17:58

Description

The current implementation of Filter on RecordBatch, Table does redundant analysis of the filter array. It would be more efficient in most cases (i.e. whenever there are multiple columns) to convert the boolean array into a selection vector and then use Take

Issue Links

• is a child of: ARROW-8894 [C++] C++ array kernels framework and execution buildout (umbrella issue)
• is related to: ARROW-5760 [C++] Optimize Take implementation
• links to: GitHub Pull Request #7475