While the current performance is good for the use case of 6.0, there are several things that can be cleaned up to make the loop scanning the index more efficient. For example: using the SortedDataInterface::Cursor directly instead of with DataThrottle; making sure we're not doing extra copies; using simpler data structures, etc.
- is related to
-
SERVER-61159 Check for duplicates before converting a non-unique index to unique
- Closed