[SPARK-55802][SQL] Fix integer overflow when computing Arrow batch bytes#54584
Closed
viirya wants to merge 1 commit intoapache:masterfrom
Closed
[SPARK-55802][SQL] Fix integer overflow when computing Arrow batch bytes#54584viirya wants to merge 1 commit intoapache:masterfrom
viirya wants to merge 1 commit intoapache:masterfrom
Conversation
HyukjinKwon
approved these changes
Mar 3, 2026
Member
Author
|
This issue is reported by @sunchao. |
sunchao
approved these changes
Mar 3, 2026
Member
|
Thanks @viirya ! |
zhengruifeng
approved these changes
Mar 3, 2026
viirya
force-pushed
the
fix-arrow-batch-bytes-overflow
branch
from
6a9089f to
edca3d5
Compare
yaooqinn
approved these changes
Mar 3, 2026
viirya
force-pushed
the
fix-arrow-batch-bytes-overflow
branch
4 times, most recently
from
8208dfb to
36e6738
Compare
Yicong-Huang
approved these changes
Mar 3, 2026
viirya
force-pushed
the
fix-arrow-batch-bytes-overflow
branch
from
36e6738 to
0de3d67
Compare
`ArrowWriter.sizeInBytes()` and `SliceBytesArrowOutputProcessorImpl .getBatchBytes()` both accumulated per-column buffer sizes (each an `Int`) into an `Int` accumulator. When the total exceeds 2 GB the sum silently wraps negative, causing the byte-limit checks controlled by `spark.sql.execution.arrow.maxBytesPerBatch` and `spark.sql.execution.arrow.maxBytesPerOutputBatch` to behave incorrectly and potentially allow oversized batches through. Fix by changing both accumulators and return types to `Long`. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
viirya
force-pushed
the
fix-arrow-batch-bytes-overflow
branch
from
0de3d67 to
076746f
Compare
viirya
added a commit
that referenced
this pull request
Mar 4, 2026
### What changes were proposed in this pull request? `ArrowWriter.sizeInBytes()` and `SliceBytesArrowOutputProcessorImpl .getBatchBytes()` both accumulated per-column buffer sizes (each an `Int`) into an `Int` accumulator. When the total exceeds 2 GB the sum silently wraps negative, causing the byte-limit checks controlled by `spark.sql.execution.arrow.maxBytesPerBatch` and `spark.sql.execution.arrow.maxBytesPerOutputBatch` to behave incorrectly and potentially allow oversized batches through. Fix by changing both accumulators and return types to `Long`. ### Why are the changes needed? Fix possible overflow when calculating Arrow batch bytes. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Existing tests. ### Was this patch authored or co-authored using generative AI tooling? Generated-by: Claude Sonnet 4.6 <noreplyanthropic.com> Closes #54584 from viirya/fix-arrow-batch-bytes-overflow. Authored-by: Liang-Chi Hsieh <viirya@gmail.com> Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com> (cherry picked from commit df195ac) Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
Member
Author
|
Merged to master/4.1. Manually backport to 4.0 in #54624 because some conflicts. Thanks @HyukjinKwon @sunchao @zhengruifeng @dongjoon-hyun @yaooqinn @Yicong-Huang |
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
7 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.
What changes were proposed in this pull request?
ArrowWriter.sizeInBytes()andSliceBytesArrowOutputProcessorImpl .getBatchBytes()both accumulated per-column buffer sizes (each anInt) into anIntaccumulator. When the total exceeds 2 GB the sum silently wraps negative, causing the byte-limit checks controlled byspark.sql.execution.arrow.maxBytesPerBatchandspark.sql.execution.arrow.maxBytesPerOutputBatchto behave incorrectly and potentially allow oversized batches through.Fix by changing both accumulators and return types to
Long.Why are the changes needed?
Fix possible overflow when calculating Arrow batch bytes.
Does this PR introduce any user-facing change?
No
How was this patch tested?
Existing tests.
Was this patch authored or co-authored using generative AI tooling?
Generated-by: Claude Sonnet 4.6 noreply@anthropic.com