-
Type: Task
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Fully Compatible
-
Query Optimization 2021-04-19, Query Optimization 2021-05-03
Sharding support for time-series collections requires using an index with simple collation. Currently queries on a view must use the same collation as the view's default collation, however after SERVER-55653 queries on time-series collections will also be able to specify a simple collation (even if the default collation is not simple) to allow for sharded time-series collections. We must make sure that our query rewrites (ie. predicate mapping to the buckets' control.min/max fields, pushing meta predicates past $_internalUnpackBucket) will return correct results in the event that the query's collation is simple and does not match the default collection collation.
Note that the optimization work for this ensures for correct rewrites for any specified collation, not just the simple collation.
- is depended on by
-
SERVER-55653 Queries on time-series views do not accept collations on indexes
- Closed
- is related to
-
SERVER-56264 Investigate if reads on time-series collections work with collation
- Closed
- related to
-
SERVER-54597 Allow users to specify non-default collation when querying on time-series collections
- Closed