- 
    Type:Improvement 
- 
    Resolution: Won't Do
- 
    Priority:Major - P3 
- 
    None
- 
    Affects Version/s: None
- 
    Component/s: None
- 
        Query Optimization
- 
        None
- 
        None
- 
        None
- 
        None
- 
        None
- 
        None
- 
        None
We found that the collection size significantly impacts the cost. The incremental cost of a fetch almost doubled when using a collection size of 1M versus 100K. We believe this is due to the same reason as for IXSCAN seeks: larger collection -> deeper B+ tree -> more expensive fetch.
As the different in cost can be huge, we should take collection size into account. We know the size of the collection inside CBR. It is in CardinalityEstimator::_collCard which should be easy to propagate to CostEstimator.