-
Notifications
You must be signed in to change notification settings - Fork 5.9k
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
util/ranger: fix wrong range calculation of prefix index when appending ranges to point ranges (#26066) #26261
util/ranger: fix wrong range calculation of prefix index when appending ranges to point ranges (#26066) #26261
Conversation
Signed-off-by: ti-srebot <[email protected]>
[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. |
/run-all-tests |
@xuyifangreeneyes you're already a collaborator in bot's repo. |
/merge |
This pull request has been accepted and is ready to merge. Commit hash: 59f75cc
|
@ti-srebot: Your PR was out of date, I have automatically updated it for you. At the same time I will also trigger all tests for you: /run-all-tests If the CI test fails, you just re-trigger the test that failed and the bot will merge the PR for you after the CI passes. 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. |
cherry-pick #26066 to release-5.0
You can switch your code base to this Pull Request by using git-extras:
# In tidb repo: git pr https://github.com/pingcap/tidb/pull/26261
After apply modifications, you can push your change to this PR via:
What problem does this PR solve?
Issue Number: close #26029
Problem Summary:
Query range of prefix index is wrong, which leads to wrong query result.
What is changed and how it works?
What's Changed & How it Works:
When building ranges for eq/in conditions, each range is a single point. If the index contains the prefix of some column, the ranges need be cut to the prefix on that column. After cutting on that column, it is possible to merge some overlapping or adjacent ranges so we call
UnionRanges
to do that. After callingUnionRanges
, two adjacent single point ranges may be merged to one range. However, we still think the ranges consist of single points and use the ranges when callingappendRanges2PointRanges
, which leads to wrong range calculation. For example, if the condition iscol1 = "xxxxxx" and col2 in (2, 3) and col3 != 4
and the index is(col1(3), col2, col3)
, we build["xxxxxx" 2, "xxxxxx" 2], ["xxxxxx" 3, "xxxxxx" 3]
from eq/in conditions. After cutting on the first column, the ranges are["xxx" 2, "xxx" 2], ["xxx" 3, "xxx" 3]
. Then, after callingUnionRanges
withrangeDetacher.mergeConsecutive = true
, the ranges are["xxx" 2, "xxx" 3]
. Later we callappendRanges2PointRanges
, the ranges are["xxx" 2 -inf, "xxx" 3 4), ("xxx" 2 4, "xxx" 3 +inf]
, which are wrong.In order to correct the wrong range calculation, we get
pointRanges
by callingrangeDetacher.buildCNFIndexRange
withrangeDetacher.mergeConsecutive = false
. In this way, two adjacent single point ranges would not be merged to one range inUnionRanges
. SopointRanges
consist of single points and we usepointRanges
when callingappendRanges2PointRanges
.Related changes
pingcap/docs
/pingcap/docs-cn
:Check List
Tests
Side effects
Release note