Skip to content

Enable arrow-ipc/zstd in datasource-arrow to make test_spill_compression pass in every config#21504

Open
AdamGS wants to merge 1 commit intoapache:mainfrom
AdamGS:adamg/issue-21503
Open

Enable arrow-ipc/zstd in datasource-arrow to make test_spill_compression pass in every config#21504
AdamGS wants to merge 1 commit intoapache:mainfrom
AdamGS:adamg/issue-21503

Conversation

@AdamGS
Copy link
Copy Markdown
Contributor

@AdamGS AdamGS commented Apr 9, 2026

Which issue does this PR close?

Rationale for this change

The spill manager assumes that all available compressions are actually available, which currently relies on feature unification with datafusion-datasource-arrow.

What changes are included in this PR?

Move arrow-ipc/zstd feature from datafusion-datasource-arrow to the workspace dependency, like lz4.

Are these changes tested?

Existing tests cover the functionality, tested individual crates locally.

Are there any user-facing changes?

Shifts some features around which might change behavior for users for DataFusion, not sure how the project reasons about this level of changes.

@github-actions github-actions bot added the datasource Changes to the datasource crate label Apr 9, 2026
@AdamGS AdamGS changed the title Enable arrow-ipc/zstd in datasource-arrow Enable arrow-ipc/zstd in datasource-arrow to make test_spill_compression pass in every config Apr 9, 2026
Copy link
Copy Markdown
Contributor

@metegenez metegenez left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

compression = [] feature becomes dead. It exists, takes no inputs, and does nothing. We should leave a comment saying why it's kept (backcompat for downstream users who wrote features = ["compression"] in their own Cargo.toml).

For architectural point of view, The spill manager still assumes "all codecs are available" without actually checking. can be the problem as well? Is spill manager's codec assumption worth a follow-up issue?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

datasource Changes to the datasource crate

Projects

None yet

Development

Successfully merging this pull request may close these issues.

test_spill_compression fails with default features

2 participants