Fix assertion failure in getNdvBySegHeapTuple for empty partitions.#1599
Open
zhangwenchao-123 wants to merge 2 commits intoapache:mainfrom
Open
Fix assertion failure in getNdvBySegHeapTuple for empty partitions.#1599zhangwenchao-123 wants to merge 2 commits intoapache:mainfrom
zhangwenchao-123 wants to merge 2 commits intoapache:mainfrom
Conversation
zhangwenchao-123
force-pushed
the
fix_analyze_core
branch
from
350241c to
c5f23fd
Compare
my-ship-it
reviewed
Mar 4, 2026
zhangwenchao-123
force-pushed
the
fix_analyze_core
branch
3 times, most recently
from
dd82ba3 to
b96c056
Compare
gfphoenix78
reviewed
Mar 5, 2026
| break; | ||
| } | ||
| } | ||
| return valid; |
Contributor
There was a problem hiding this comment.
All of the ndvbsSlots could be release if the valid becomes false. On the other side, the ndvbsSlots and its pointer array could be also freed in aggregate_leaf_partition_ndvbs no matter it's valid or not.
Contributor
Author
There was a problem hiding this comment.
Right, aggregate_leaf_partition_ndvbs will free all ndvbsSlots regardless of validity. The early cleanup here for empty partitions (setting slot to NULL after free) is just to keep the slot array consistent during iteration — it won't cause a double-free since the cleanup function checks for NULL.
yjhjstz
reviewed
Mar 9, 2026
When merging leaf partition statistics for a partitioned table, getNdvBySegHeapTuple could hit Assert(valuetype == FLOAT8OID) if a partition had relTuples == 0 and its pg_statistic entry lacked a valid STATISTIC_KIND_NDV_BY_SEGMENTS slot (or the slot had an unexpected element type). The original guard condition only handled two cases: 1. Non-empty partition with NDV value == 0 2. Non-empty partition without NDV_BY_SEGMENTS It missed the case where an empty partition (relTuples == 0) has a pg_statistic entry but no valid FLOAT8OID NDV_BY_SEGMENTS slot, causing the code to fall through to the assertion. Fix by checking valuetype != FLOAT8OID upfront: skip empty partitions gracefully, and mark non-empty partitions as invalid. Authored-by: Zhang Wenchao zwcpostgres@gmail.com
- Add back Assert(valuetype == FLOAT8OID) as defensive check after the type guard (per gfphoenix78 review) - Replace Assert(nvalues == 1) with runtime validity check to avoid assertion failure in non-debug builds (per yjhjstz review) Authored-by: Zhang Wenchao <zhangwenchao@apache.org>
zhangwenchao-123
force-pushed
the
fix_analyze_core
branch
from
31b9bcf to
080c758
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
When merging leaf partition statistics for a partitioned table, getNdvBySegHeapTuple could hit Assert(valuetype == FLOAT8OID) if a partition had relTuples == 0 and its pg_statistic entry lacked a valid STATISTIC_KIND_NDV_BY_SEGMENTS slot (or the slot had an unexpected element type).
The original guard condition only handled two cases:
Fix by checking valuetype != FLOAT8OID upfront: skip empty partitions gracefully, and mark non-empty partitions as invalid.
Fixes #ISSUE_Number
What does this PR do?
Type of Change
Breaking Changes
Test Plan
make installcheckmake -C src/test installcheck-cbdb-parallelImpact
Performance:
User-facing changes:
Dependencies:
Checklist
Additional Context
CI Skip Instructions