Arrow: Release vectors decoded from dictionary encoded columns - #17742
Open
chiragkava wants to merge 1 commit into
Open
Arrow: Release vectors decoded from dictionary encoded columns#17742chiragkava wants to merge 1 commit into
chiragkava wants to merge 1 commit into
Conversation
chiragkava
force-pushed
the
fix-dict-encoded-vector-leak
branch
from
August 20, 2026 11:11
aded587 to
fc6250f
Compare
anoopj
approved these changes
Aug 20, 2026
| } | ||
|
|
||
| ColumnVector[] columnVectors = new ColumnVector[readers.length]; | ||
| // release the decoded vectors the previous batch materialized, safe with reused containers |
Member
There was a problem hiding this comment.
This comment is not very readable.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #17722.
ColumnVector.getArrowVector()allocates a decoded vector per call for dictionary encoded columns and nothing released it (details in the issue). Following the ownership model of #17296, the decoded vector is now materialized once per batch andArrowBatchReaderreleases it when the nextread()invalidates the batch and onclose(). Only dict-decoded vectors are released, soreuseContainersand the reader-owned holder vectors are untouched.Spark/Flink don't use this code path.
Tested with a differential pair in
TestArrowReader: the materializing test fails on main (33,280 bytes retained for 100 rows / 1 batch / 1 string column) and passes with the fix. the non-materializing control passes on both.This contribution was developed with AI assistance Claude Code (Opus 5) and I have reviewed it