Uh oh!
There was an error while loading. Please reload this page.
fix: use spill writer's schema instead of the first batch schema for spill files - #21293
Merged
Merged
Conversation
gruuyaforce-pushed
the
union-nullable-spill-fix
branch
from
April 1, 2026 08:58
debff98 to
fd58942Comparegruuyaforce-pushed
the
union-nullable-spill-fix
branch
3 times, most recently
from
April 1, 2026 11:19
d0bcb4e to
2f08c97Comparealamb
approved these changes
Apr 2, 2026
| } | ||
| } | ||
| #[cfg(test)] |
Contributor
There was a problem hiding this comment.
I wonder if the unit tests are needed give the end to end tests?
ContributorAuthor
There was a problem hiding this comment.
I was actually of the opposite opinion—the e2e test is slightly unwieldy to me, and convoluted in a way, since it tries to reproduce the bug using non-standard primitives and hit the right spill conditions in an unnatural way (because otherwise it's hard to repro). For instance union/repartition exec used directly instead of sql/dataframe API, special task ctx, sort_batch called directly on batches instead of wrapping with SortExec.
Also InProgressSpillFile spill file unit tests were missing anyway.
Either way, I'm fine with dropping either one (or none) of those, let me know what you think.
Contributor
There was a problem hiding this comment.
I think this it is fine to keep both
gruuyaforce-pushed
the
union-nullable-spill-fix
branch
from
April 3, 2026 07:53
2f08c97 to
8fbd95eComparealamb
approved these changes
Apr 3, 2026
| @@ -0,0 +1,162 @@ | |||
| // Licensed to the Apache Software Foundation (ASF) under one | |||
Contributor
There was a problem hiding this comment.
Could we please not add a new test binary? Instead add it to the existing core_integration test?
Each new test binary takes like 100MB of disk space on the runner
ContributorAuthor
There was a problem hiding this comment.
Oh for sure, give me a minute.
ContributorAuthor
There was a problem hiding this comment.
Done, moved to memory_limit dir as there's another spill-themed test there already.
gruuyaforce-pushed
the
union-nullable-spill-fix
branch
from
April 3, 2026 20:57
8fbd95e to
4480b84Comparegruuyaforce-pushed
the
union-nullable-spill-fix
branch
from
April 3, 2026 21:08
4480b84 to
e133dd3Comparealamb
commented
Apr 3, 2026
Contributor
Thanks again @gruuya |
Uh oh!
There was an error while loading. Please reload this page.
This was referenced Apr 4, 2026
alamb pushed a commit
to alamb/datafusion
that referenced
this pull request
Apr 6, 2026
…spill files (apache#21293) <!-- We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closesapache#123` indicates that this PR will close issue apache#123. --> - Closesapache#21292. <!-- Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed. Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes. --> `InProgressSpillFile` will use the first batch schema when it has no writer, which could have different nullability from the subsequent batches (e.g. due to union-ing a literal projection and a table column projection). This can then lead to a panic in `sort_batch`. `InProgressSpillFile` already has access to the canonical schema that the spill file should have (`self.spill_writer.schema()`). <!-- There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR. --> Single line fix: instead of using the first batch schema for the spill file schema, use the spill_writer's schema instead. The rest of the changes are two new tests. <!-- We typically require tests for all PRs in order to: 1. Prevent the code from being accidentally broken by subsequent changes 2. Serve as another way to document the expected behavior of the code If tests are not included in your PR, please explain why (for example, are they covered by existing tests)? --> Yes, two new tests are added. <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. --> <!-- If there are any breaking changes to public APIs, please add the `api change` label. -->
alamb added a commit
that referenced
this pull request
Apr 7, 2026
Contributor
FYI I backported this to 52 and we will do it for 53 as well |
lizardoluis pushed a commit
to splitgraph/arrow-datafusion
that referenced
this pull request
Apr 16, 2026
…spill files (apache#21293) ## Which issue does this PR close? <!-- We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closesapache#123` indicates that this PR will close issue apache#123. --> - Closesapache#21292. ## Rationale for this change <!-- Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed. Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes. --> `InProgressSpillFile` will use the first batch schema when it has no writer, which could have different nullability from the subsequent batches (e.g. due to union-ing a literal projection and a table column projection). This can then lead to a panic in `sort_batch`. `InProgressSpillFile` already has access to the canonical schema that the spill file should have (`self.spill_writer.schema()`). ## What changes are included in this PR? <!-- There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR. --> Single line fix: instead of using the first batch schema for the spill file schema, use the spill_writer's schema instead. The rest of the changes are two new tests. ## Are these changes tested? <!-- We typically require tests for all PRs in order to: 1. Prevent the code from being accidentally broken by subsequent changes 2. Serve as another way to document the expected behavior of the code If tests are not included in your PR, please explain why (for example, are they covered by existing tests)? --> Yes, two new tests are added. ## Are there any user-facing changes? <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. --> <!-- If there are any breaking changes to public APIs, please add the `api change` label. -->
Rich-T-kid pushed a commit
to Rich-T-kid/datafusion
that referenced
this pull request
Apr 21, 2026
…spill files (apache#21293) ## Which issue does this PR close? <!-- We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closesapache#123` indicates that this PR will close issue apache#123. --> - Closesapache#21292. ## Rationale for this change <!-- Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed. Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes. --> `InProgressSpillFile` will use the first batch schema when it has no writer, which could have different nullability from the subsequent batches (e.g. due to union-ing a literal projection and a table column projection). This can then lead to a panic in `sort_batch`. `InProgressSpillFile` already has access to the canonical schema that the spill file should have (`self.spill_writer.schema()`). ## What changes are included in this PR? <!-- There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR. --> Single line fix: instead of using the first batch schema for the spill file schema, use the spill_writer's schema instead. The rest of the changes are two new tests. ## Are these changes tested? <!-- We typically require tests for all PRs in order to: 1. Prevent the code from being accidentally broken by subsequent changes 2. Serve as another way to document the expected behavior of the code If tests are not included in your PR, please explain why (for example, are they covered by existing tests)? --> Yes, two new tests are added. ## Are there any user-facing changes? <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. --> <!-- If there are any breaking changes to public APIs, please add the `api change` label. -->
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 freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
Which issue does this PR close?
Rationale for this change
InProgressSpillFilewill use the first batch schema when it has no writer, which could have different nullability from the subsequent batches (e.g. due to union-ing a literal projection and a table column projection). This can then lead to a panic insort_batch.InProgressSpillFilealready has access to the canonical schema that the spill file should have (self.spill_writer.schema()).What changes are included in this PR?
Single line fix: instead of using the first batch schema for the spill file schema, use the spill_writer's schema instead.
The rest of the changes are two new tests.
Are these changes tested?
Yes, two new tests are added.
Are there any user-facing changes?