[ARROW-12659] [C++][Compute] Support SimplifyWithGuarantee(is_null(foo), is_valid(foo))

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 8.0.0
• Component/s: C++
• Labels: dataset datasets expression expressions pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/28408
• Assignee: Ben Kietzman
• Reporter: David Li
• Created: 05/May/21 18:06
• Updated: 11/Jan/23 08:27
• Resolved: 21/Apr/22 18:55

Description

Simplifying a predicate using is_null with a guarantee using invert(is_null()) doesn't lead to any simplification.

Once supported, tests should be added for Datasets to exercise row group filtering (and to generate the necessary clauses from statistics).

Issue Links

• blocks: ARROW-13848 [C++] and() in a dataset filter
• is related to: ARROW-15312 [R][C++] filtering a Parquet dataset with is.na() misses some rows
• is related to: ARROW-11762 [C++][Dataset] Refactor Partitioning to explicitly treat null and absent fields identically
• links to: GitHub Pull Request #10253
• links to: GitHub Pull Request #12891