[IMPALA-10709] Min/max filters should be enabled for joins on sorted columns in Parquet tables

• Type: Test
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: Impala 4.1.0
• Epic Color: ghx-label-7
• Assignee: Qifan Chen
• Reporter: Qifan Chen
• Created: 19/May/21 17:51
• Updated: 12/Jul/21 15:07
• Resolved: 12/Jul/21 15:07

Description

Currently, the min/max filter feature is turned off by default (MINMAX_FILTER_THRESHOLD=0).

When joining into sorted columns in a fact Parquet table created by Imoala, the feature can be turned on by default. This is because Impala sorts the data in sort by columns in each data file during population. A min/max filter can be used to easily reject pages not overlapping with the search region specified in the filter.