-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: 6.0.0, 6.1.0, 6.2.0-rc4
-
Component/s: Query Planning
-
None
-
Storage Execution
-
Fully Compatible
-
ALL
-
v6.3, v6.0
-
Execution Team 2023-02-20, Execution Team 2023-03-06
When a clustered collection has a non-"simple" default collation, its _id values are encoded using the collection's default collation. This means the minRecord and maxRecord bounds cannot be constrained when querying by _id and using a collation differing from the collection's default. However, handleRIDRangeScan() incorrectly always permits tighter bounds when using {locale: "simple"}.
- is caused by
-
SERVER-61260 Support collations on clustered collections
- Closed
- is depended on by
-
SERVER-72435 Add more targeted collation testing for updateOne, deleteOne, and findAndModifys without shard key
- Closed