[ARROW-10556] [C++] Caching pre computed data based on FunctionOptions in the kernel state

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: C++
• External issue URL: https://github.com/apache/arrow/issues/26521
• Assignee: Unassigned
• Reporter: Maarten Breddels
• Created: 11/Nov/20 10:44
• Updated: 11/Jan/23 08:14

Description

See discussion here:

https://github.com/apache/arrow/pull/8621#issuecomment-724796243

 

A kernel might need to pre-compute something based on the function options passed. Since the Kernel-FunctionOptions mapping is not 1-to-1, it does not make sense to store this in the function option object. 

Currently, match_substring calculates a `prefix_table` on each Exec call. In trim (https://github.com/apache/arrow/pull/8621) we compute a vector<bool> on each Exec call. This should be done only once and cached in the kernel state instead.

 

Issue Links

• depends upon: ARROW-10655 [C++] Add LRU cache facility
• is related to: ARROW-9128 [C++] Implement string space trimming kernels: trim, ltrim, and rtrim
• relates to: ARROW-10097 [C++] Persist SetLookupState in between usages of IsIn when filtering dataset batches