-
Notifications
You must be signed in to change notification settings - Fork 5.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
planner: fix incorrect maintenance of handleColHelper
for recursive CTE
#55732
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #55732 +/- ##
=================================================
- Coverage 72.8514% 56.1953% -16.6561%
=================================================
Files 1601 1729 +128
Lines 445302 622110 +176808
=================================================
+ Hits 324409 349597 +25188
- Misses 100894 248845 +147951
- Partials 19999 23668 +3669
Flags with carried forward coverage won't be shown. Click here to find out more.
|
Please add a description of how to handle CTE in the comments for the handleHelper attribute. |
/retest |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me. Thanks! 👍
@@ -7000,14 +7030,11 @@ func (b *PlanBuilder) buildRecursiveCTE(ctx context.Context, cte ast.ResultSetNo | |||
// Build seed part plan. | |||
saveSelect := x.SelectList.Selects | |||
x.SelectList.Selects = x.SelectList.Selects[:i] | |||
// We're rebuilding the seed part, so we pop the result we built previously. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can I ask why we need to remove this?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The popMap()
here is moved right after b.buildSelect()
and b.buildSetOpr()
in this function.
This should make it more clear and explicit that we directly drop the entries pushed by them.
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: Rustin170506, winoros The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
[LGTM Timeline notifier]Timeline:
|
/retest |
1 similar comment
/retest |
In response to a cherrypick label: new pull request created to branch |
Signed-off-by: ti-chi-bot <[email protected]>
In response to a cherrypick label: new pull request created to branch |
Signed-off-by: ti-chi-bot <[email protected]>
Signed-off-by: ti-chi-bot <[email protected]>
In response to a cherrypick label: new pull request created to branch |
Signed-off-by: ti-chi-bot <[email protected]>
In response to a cherrypick label: new pull request created to branch |
* origin/master: (33 commits) build(deps): bump github.com/prometheus/common from 0.55.0 to 0.57.0 (pingcap#55762) build(deps): bump golang.org/x/sys from 0.24.0 to 0.25.0 (pingcap#55894) planner: fix incorrect estRows with global index and json column (pingcap#55842) ddl, stat: switch to new struct for add/truncate/drop partition (pingcap#55893) planner: hide instance plan cache eviction log if no plan is evicted (pingcap#55918) expression: support tidb encode key function (pingcap#51678) planner: fix incorrect maintenance of `handleColHelper` for recursive CTE (pingcap#55732) executor: some code refine of hash join v2 (pingcap#55887) infoschema, meta: fix wrong auto id after `rename table` (pingcap#55847) ddl/ingest: set `minCommitTS` when detect remote duplicate keys (pingcap#55588) planner: move index advisor into the kernel (pingcap#55874) ddl, stats: switch to new struct for create/truncate table (pingcap#55811) executor: avoid new small objects in probe stage of hash join v2 (pingcap#55855) *: Add tidbCPU/tikvCPU into system tables (pingcap#55455) ddl: use static contexts in `NewReorgCopContext` (pingcap#55823) executor: fix index out of range bug in hash join v2 (pingcap#55864) executor: record index usage for the clustered primary keys (pingcap#55602) domain: load all non-public tables into infoschema (pingcap#55142) test: fix unstable TestShowViewPriv (pingcap#55868) ttl: add `varbinary` case for `TestSplitTTLScanRangesWithBytes` (pingcap#55863) ...
What problem does this PR solve?
Issue Number: close #55666
Problem Summary:
If the recursive part of a CTE references tables other than the CTE itself, the
b.handleHelper
won't be correctly cleared duringbuildRecursiveCTE()
.It's because this function relies on an error from
buildSelect()
orbuildSetOpr()
to recognize the boundary between the seed part and the recursive part. Butb.handleHelper
won't be cleared when an error happens.When this happens, there will be some wrong and redundant entries in
b.handleHelper
, which makes users of this field likebuildDelete()
get the wrong result or panic.What changed and how does it work?
Since this PR needs to be cherry-picked, I can't change too many.
buildRecursiveCTE()
, clear theb.handleHelper
if an error is returned frombuildSelect()
orbuildSetOpr()
.b.handleHelper
more clear for recursive CTE:tryBuildCTE()
will always push an entry into handleHelper when successful.buildRecursiveCTE()
will always push an entry into handleHelper when successful.Check List
Tests
Side effects
Documentation
Release note
Please refer to Release Notes Language Style Guide to write a quality release note.