[CALCITE-3711] Correlate should override estimateRowCount

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.21.0
• Fix Version/s: 1.22.0
• Labels: pull-request-available
• Assignee: Vladimir Sitnikov
• Reporter: Ruben Q L
• Created: 06/Jan/20 08:57
• Updated: 06/Mar/20 04:35
• Resolved: 06/Jan/20 20:33

Description

Currently, Correlate inherits the estimateRowCount implementation from AbstractRelNode:

public double estimateRowCount(RelMetadataQuery mq) {
  return 1.0;
}

which gives a "default" (but unrealistic) row count computation.
This has several issues:

• This value may be used by Correlate#computeSelfCost (and by inheritance by LogicalCorrelate / EnumerableCorrelate):

  @Override public RelOptCost computeSelfCost(final RelOptPlanner planner, final RelMetadataQuery mq) {
    double rowCount = mq.getRowCount(this);  // by default: estimateRowCount, i.e. 1.0
    ...
  

• As the current state, this value will definitely by used by several join algorithms in their cost computation if their left/right child is a Correlate:

  public class EnumerableHashJoin extends Join implements EnumerableRel {
  ...
    @Override public RelOptCost computeSelfCost(RelOptPlanner planner, RelMetadataQuery mq) {
      ...
      final double rightRowCount = right.estimateRowCount(mq); // 1.0 if right is a Correlate
      final double leftRowCount = left.estimateRowCount(mq); // 1.0 if left is a Correlate
      ... 
  

Even though cost computation is pluggable (which would help solving issues like the first one, but not issues like the second one), IMHO we should provide a more realistic default Correlate#estimateRowCount computation.

Issue Links

• links to: GitHub Pull Request #1726