[FLINK-19780] FlinkRelMdDistinctRowCount#getDistinctRowCount(Calc) will always return 0 when number of rows are large

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Implemented
• Fix Version/s: 1.12.0
• Component/s: Table SQL / Planner
• Labels: pull-request-available
• Assignee: Caizhi Weng
• Reporter: Caizhi Weng
• Created: 23/Oct/20 08:06
• Updated: 27/Oct/20 02:21
• Resolved: 27/Oct/20 02:21

Description

Due to CALCITE-4351 FlinkRelMdDistinctRowCount#getDistinctRowCount(Calc) will always return 0 when number of rows are large.

What I would suggest is to introduce our own FlinkRelMdUtil#numDistinctVals to treat small and large inputs in different ways. For small inputs we use the more precise RelMdUtil#numDistinctVals and for large inputs we copy the old, approximated implementation of RelMdUtil#numDistinctVals.

This is a temporary solution. When CALCITE-4351 is fixed we should revert this commit.

Issue Links

• is caused by: CALCITE-4351 RelMdUtil#numDistinctVals always returns 0 for large inputs
• is caused by: FLINK-16579 Upgrade Calcite version to 1.26 for Flink SQL
• links to: GitHub Pull Request #13764