MINOR: Reject non zero offset on array import - #1263
Open
robert3005 wants to merge 5 commits into
Open
Conversation
robert3005
requested review from
jbonofre,
laurentgo,
lidavidm and
wgtmac
as code owners
August 19, 2026 13:35
This comment has been minimized.
This comment has been minimized.
jbonofre
requested changes
Aug 27, 2026
Author
|
Thanks for the suggestion, I applied it and added test to make sure we don't regress it in the future |
Member
|
@robert3005 thanks! I'm doing a new pass 😄 |
Member
|
By the way, we merged a fix on CI, I will rebase. |
jbonofre
force-pushed
the
rk/importoffset
branch
from
August 30, 2026 06:49
07f209b to
20b5fea
Compare
kou
pushed a commit
that referenced
this pull request
Aug 31, 2026
The `JNI macos-15-intel x86_64` job currently fails on `main` and on every open PR while installing dependencies: ``` ==> Installing node Error: node: no bottle available! This is a Tier 3 configuration: https://docs.brew.sh/Support-Tiers#tier-3 Installing node has failed! `brew bundle` failed! 1 Brewfile dependency failed to install ``` `apache/arrow`'s `cpp/Brewfile` lists `brew "node"`, and Homebrew treats x86_64 macOS as a tier 3 configuration and no longer builds a bottle for it. `brew bundle` doesn't build from source, so the step aborts before anything is built. We don't need Node.js for the JNI build, so this skips installing it via `HOMEBREW_BUNDLE_BREW_SKIP`, the same way #1269 did for `aws-sdk-cpp` and gRPC. Failing runs for reference: * `main`: https://github.com/apache/arrow-java/actions/runs/33283736388/job/99183221397 * PR #1263: https://github.com/apache/arrow-java/actions/runs/33297775855/job/99221106761
Signed-off-by: Robert Kruszewski <github@robertk.io>
Signed-off-by: Robert Kruszewski <github@robertk.io>
jbonofre
force-pushed
the
rk/importoffset
branch
from
August 31, 2026 05:32
20b5fea to
520040f
Compare
Null arrays export no buffers (n_buffers == 0), so the offset has nothing to apply to and never enters the import: doImport short-circuits to an empty buffer list and NullVector.loadFieldBuffers only reads the field node length. Arrow C++ still propagates a slice offset onto the struct for these arrays, so rejecting them regressed imports that previously succeeded, e.g. a sliced all-null column.
Member
|
@robert3005 I updated the I will merge as soon as CI is green. |
jbonofre
approved these changes
Aug 31, 2026
lidavidm
approved these changes
Aug 31, 2026
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
3 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's Changed
Instead of silently accepting the array and pretending the import went fine reject arrays where offset is non 0. Add mention to the docs that this is not supported
This is documentation and negative case for #251