[ARROW-12673] [C++] Configure a custom handler for rows with incorrect column counts

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 6.0.0
• Component/s: C++
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/28421
• Assignee: Nate Clark
• Reporter: Nate Clark
• Created: 06/May/21 18:57
• Updated: 11/Jan/23 08:27
• Resolved: 20/Sep/21 17:57

Description

Currently when a row does not have the correct number of columns an error is generated and parsing stops. It would be better if this behavior was customizable depending on the caller so that rows could either be ignored or altered to have the correct column count.

Ideally there would be some builtin handlers for this but it would be good for the caller to be able to implement their own handler if they wanted to collect the malformed rows and report them to a different component or maybe add fields based on the existing fields in the row.

Issue Links

• relates to: ARROW-12001 [C++][CSV] Allow missing columns at end of row
• links to: GitHub Pull Request #10202
• links to: GitHub Pull Request #10790