planner: Add control flag to keep or remove ORDER BY in subquery #33173
Conversation
|
[REVIEW NOTIFICATION] This pull request has been approved by:
To complete the pull request process, please ask the reviewers in the list to review by filling The full list of commands accepted by this bot can be found here. Reviewer can indicate their review by submitting an approval review. |
ti-chi-bot
added
the
release-note
|
|
ti-chi-bot
added
needs-rebase
|
Welcome @yuanhsh! |
|
Code Coverage Details: https://codecov.io/github/pingcap/tidb/commit/7bc1164629562c1fea0c477ca799c1727ed7dda6 |
ti-chi-bot
removed
the
needs-rebase
ti-chi-bot
added
size/L
ti-chi-bot
added
size/M
| {Scope: ScopeGlobal | ScopeSession, Name: TiDBRemoveOrderbyInSubquery, Value: BoolToOnOff(DefTiDBRemoveOrderbyInSubquery), Type: TypeBool, SetSession: func(s *SessionVars, val string) error { | ||
| s.RemoveOrderbyInSubquery = TiDBOptOn(val) | ||
| return nil | ||
| }}, |
There was a problem hiding this comment.
I think it's not a global variable in this way. Mark the ScopeGlobal will enable the persistence to GLOBAL_VARIABLES table while the value persisted is never inherited by session (You use DefTiDBRemoveOrderbyInSubquery to init s. RemoveOrderbyInSubquery).
You can just remove ScopeGlobal or implement the full ScopeGlobal logic.
There was a problem hiding this comment.
Thanks, I have removed ScopeGlobal as suggested.
There was a problem hiding this comment.
I think this looked right to me, but I'm not sure because of the force-push.
The Value: only refers to the default value, once it is persisted it will use what is stored internally. SetSession is called both when a session value is changed, and on init for new sessions (on init the val will be the global value, which matches the GLOBAL_VARIABLES -- although we have a cache of it internally called sysvar_cache).
There was a problem hiding this comment.
The downside of ScopeSession only is that it doesn't persist, so each time a user wants this optimization to apply -- they'll need to SET sysvar = x. Which limits its appeal.
There was a problem hiding this comment.
You are right, my fault, I mixed the situation, @yuanhsh please ignore my previous comment.
There was a problem hiding this comment.
OK, I will revert to previous version.
@morgo Thank you for advising!
We introduced a new contral flag "tidb_remove_orderby_in_subquery" to indicate whether to remove ORDER BY clause in subquery if the subquery doesn't have LIMIT clause. The default value is false, which means no behaviour change in current system, unless users explicitly change this setting.
|
@chrysan: Thanks for your review. The bot only counts approvals from reviewers and higher roles in list, but you're still welcome to leave your comments. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the ti-community-infra/tichi repository. |
ti-chi-bot
added
status/LGT2
|
/merge |
|
This pull request has been accepted and is ready to merge. Commit hash: 7bc1164
|
ti-chi-bot
added
the
status/can-merge
|
/run-mysql-test |
What problem does this PR solve?
Issue Number: close #32900
Unnecessary OrderBy not removed resulting in poor performance.
Problem Summary:
What is changed and how it works?
We introduced a new control flag "tidb_remove_orderby_in_subquery" to indicate
whether to remove ORDER BY clause in subquery if the subquery doesn't have LIMIT
clause. The default value is OFF, which means no behavior change in current
system, unless users explicitly change this setting.
Check List
Tests
Side effects
Documentation
Release note