[SPARK-42685] optimize byteToString routines

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 3.5.0
• Fix Version/s: 3.5.0
• Component/s: Spark Core
• Assignee: Unassigned
• Reporter: Alkis Evlogimenos
• Created: 06/Mar/23 16:06
• Updated: 10/Mar/23 18:25
• Resolved: 10/Mar/23 17:28

Description

Utils.byteToString routines are slow because they use BigInt and BigDecimal. This is causing visible CPU usage (1-2% in scan benchmarks).

Issue Links

• links to: [Github] Pull Request #40301 (alkis)
• links to: [Github] Pull Request #40301 (alkis)