[ARROW-9863] [C++] [PARQUET] Optimize meta data recovery of ApplicationVersion

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 1.0.0
• Fix Version/s: 2.0.0
• Component/s: C++
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/25900
• Assignee: Antoine Pitrou
• Reporter: Pierre Belzile
• Created: 26/Aug/20 13:22
• Updated: 11/Jan/23 08:09
• Resolved: 03/Sep/20 15:56

Description

The class contains two large regexes which are compiled in the ApplicationVersion::ApplicationVersion(const std:::string) constructor. This is the constructor that is used when reading files. 

I stopped a server in gdb that had been processing several files at once and 4 threads out of 8 were building those regexes!

Issue Links

• links to: GitHub Pull Request #8107