Mattias Jonsson
Mattias Jonsson
With this diff, partitioning pruning work and QPS is even a few percentage higher than the non-partitioned table instead: ```diff diff --git a/pkg/planner/core/rule_partition_processor.go b/pkg/planner/core/rule_partition_processor.go index 1d9e3a0708..092ed936eb 100644 --- a/pkg/planner/core/rule_partition_processor.go +++...
Some investigation results: The removed code was added in #22452 and then in #22953 a related changed required to only allow a single column in the partitioning expression for pruning...
Any tips on how to create a test for this? EXPLAIN FOR CONNECTION will not show partitioning pruning (#55669).
regression from #53544
Should DM supporting MySQL and MariaDB Vector data type be a part of this as well, or should that be a separate issue?
> @mjonss Thanks for your contribution. I am currently analyzing the rewrite branch and I am interested about your change: Ultimately, what does having this extra index handled by mysql_random_data_load...