[SPARK-53342][SQL] Fix Arrow converter to handle multiple record batches in single IPC stream #52090
+538
−3
Conversation
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
grundprinzip
force-pushed
the
SPARK-53342
branch
from
89bf987 to
ac1f23b
Compare
HyukjinKwon
approved these changes
Aug 25, 2025
|
I would like @hvanhovell to take a look though if he finds some time. |
| s"Expected ${inputRows.length} rows processed, got ${iterator.totalRowsProcessed}") | ||
| } | ||
|
|
||
| test("multiple record batches in single IPC stream") { |
There was a problem hiding this comment.
Thank you for adding this test coverage.
dongjoon-hyun
approved these changes
Aug 26, 2025
|
Merged to master. Could you make backporting PRs for all live release branches, please, @grundprinzip ? |
grundprinzip
added a commit
to grundprinzip/spark
that referenced
this pull request
Aug 26, 2025
…hes in single IPC stream ### What changes were proposed in this pull request? This PR adds a new method in ArrowConverters that allows properly decoding an Arrow IPC stream, which can contain multiple record batches. All of the other methods can only deal with message streams that contain exactly one record batch. ### Why are the changes needed? Previously, when an Arrow IPC stream contained multiple record batches, only the first batch would be processed and the remaining batches would be ignored. This resulted in data loss and incorrect results when working with Arrow data that was serialized as a single stream with multiple batches. ### Does this PR introduce _any_ user-facing change? Yes. This fixes a data correctness issue where users would lose data when processing Arrow streams with multiple batches. The behavior change is that all batches in a stream are now correctly processed instead of only the first one. ### How was this patch tested? Added comprehensive test cases. ### Was this patch authored or co-authored using generative AI tooling? Tests Generated-by: Claude Code 🤖 Generated with [Claude Code](https://claude.ai/code) Closes apache#52090 from grundprinzip/SPARK-53342. Authored-by: Martin Grund <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
grundprinzip
added a commit
to grundprinzip/spark
that referenced
this pull request
Aug 26, 2025
…hes in single IPC stream ### What changes were proposed in this pull request? This PR adds a new method in ArrowConverters that allows properly decoding an Arrow IPC stream, which can contain multiple record batches. All of the other methods can only deal with message streams that contain exactly one record batch. ### Why are the changes needed? Previously, when an Arrow IPC stream contained multiple record batches, only the first batch would be processed and the remaining batches would be ignored. This resulted in data loss and incorrect results when working with Arrow data that was serialized as a single stream with multiple batches. ### Does this PR introduce _any_ user-facing change? Yes. This fixes a data correctness issue where users would lose data when processing Arrow streams with multiple batches. The behavior change is that all batches in a stream are now correctly processed instead of only the first one. ### How was this patch tested? Added comprehensive test cases. ### Was this patch authored or co-authored using generative AI tooling? Tests Generated-by: Claude Code 🤖 Generated with [Claude Code](https://claude.ai/code) Closes apache#52090 from grundprinzip/SPARK-53342. Authored-by: Martin Grund <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
grundprinzip
added a commit
to grundprinzip/spark
that referenced
this pull request
Aug 26, 2025
…hes in single IPC stream This PR adds a new method in ArrowConverters that allows properly decoding an Arrow IPC stream, which can contain multiple record batches. All of the other methods can only deal with message streams that contain exactly one record batch. Previously, when an Arrow IPC stream contained multiple record batches, only the first batch would be processed and the remaining batches would be ignored. This resulted in data loss and incorrect results when working with Arrow data that was serialized as a single stream with multiple batches. Yes. This fixes a data correctness issue where users would lose data when processing Arrow streams with multiple batches. The behavior change is that all batches in a stream are now correctly processed instead of only the first one. Added comprehensive test cases. Tests Generated-by: Claude Code 🤖 Generated with [Claude Code](https://claude.ai/code) Closes apache#52090 from grundprinzip/SPARK-53342. Authored-by: Martin Grund <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
|
Thank you, @grundprinzip . |
|
Oh, we cannot touch |
|
Thanks! |
dongjoon-hyun
pushed a commit
that referenced
this pull request
Aug 26, 2025
… batches in single IPC stream ### What changes were proposed in this pull request? Cherry-Pick: #52090 This PR adds a new method in ArrowConverters that allows properly decoding an Arrow IPC stream, which can contain multiple record batches. All of the other methods can only deal with message streams that contain exactly one record batch. ### Why are the changes needed? Previously, when an Arrow IPC stream contained multiple record batches, only the first batch would be processed and the remaining batches would be ignored. This resulted in data loss and incorrect results when working with Arrow data that was serialized as a single stream with multiple batches. ### Does this PR introduce _any_ user-facing change? Yes. This fixes a data correctness issue where users would lose data when processing Arrow streams with multiple batches. The behavior change is that all batches in a stream are now correctly processed instead of only the first one. ### How was this patch tested? Added comprehensive test cases. ### Was this patch authored or co-authored using generative AI tooling? Tests Generated-by: Claude Code 🤖 Generated with [Claude Code](https://claude.ai/code) Closes #52090 from grundprinzip/SPARK-53342. Authored-by: Martin Grund <martin.grunddatabricks.com> Closes #52130 from grundprinzip/SPARK-53342-4.0. Authored-by: Martin Grund <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
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.
What changes were proposed in this pull request?
This PR adds a new method in ArrowConverters that allows properly decoding an Arrow IPC stream, which can contain multiple record batches. All of the other methods can only deal with message streams that contain exactly one record batch.
Why are the changes needed?
Previously, when an Arrow IPC stream contained multiple record batches, only the first batch would be processed and the remaining batches would be ignored. This resulted in data loss and incorrect results when working with Arrow data that was serialized as a single stream with multiple batches.
Does this PR introduce any user-facing change?
Yes. This fixes a data correctness issue where users would lose data when processing Arrow streams with multiple batches. The behavior change is that all batches in a stream are now correctly processed instead of only the first one.
How was this patch tested?
Added comprehensive test cases.
Was this patch authored or co-authored using generative AI tooling?
Tests Generated-by: Claude Code
🤖 Generated with Claude Code