[HIVE-4727] Optimize ORC StringTreeReader::nextVector to not create dictionary of strings for each call to nextVector

• Type: Sub-task
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: vectorization-branch, 0.13.0
• Assignee: Sarvesh Sakalanaga
• Reporter: Sarvesh Sakalanaga
• Created: 13/Jun/13 01:05
• Updated: 23/Oct/13 21:59
• Resolved: 13/Jun/13 19:38

Description

Currently ORC StringTreeReader::nextVector creates dictionary of strings for each call to nextVector. This leads to bad perf as there is huge memory allocation and deallocation on each call. Since the dictionary does not change within a stripe, StringTreeReader::nextVector should be optimized to create this dictionary only on stripe read.