feat: implement native S3 write support#4547
Open
kazantsev-maksim wants to merge 58 commits into
Open
Conversation
This reverts commit 768b3e9.
# Conflicts: # native/Cargo.lock # native/core/Cargo.toml
kazantsev-maksim
changed the title
Contributor
Author
|
@andygrove @mbutrovich @comphead Could you please give feedback on these changes – do they make sense to you? |
Contributor
|
Are there any limitations from doing it with |
comphead
reviewed
Jun 8, 2026
comphead
left a comment
comphead
left a comment
Contributor
There was a problem hiding this comment.
Thanks @kazantsev-maksim I think we need to proceed with #3209 first to make sure writer works properly with Spark tests. I'll prioritize it this week!
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Which issue does this PR close?
Part of: #1625
Rationale for this change
Currently, when Comet executes ETL queries that read from Parquet, perform a transformation, and then write back to Parquet (or S3-backed Parquet), a columnar-to-row conversion is required before the write step, because the write path falls back to the JVM Spark writer. This conversion adds unnecessary overhead and negates the performance benefits of native execution.
What changes are included in this PR?
native/core/src/execution/operators/parquet_writer.rs — Extended the native Parquet writer to support writing to S3-compatible object storage. Added S3 object store registration and wired it into the DataFusion execution context so that output paths with s3:// / s3a:// schemes are handled natively via the object_store crate.
spark/src/main/scala/org/apache/comet/serde/operator/CometDataWritingCommand.scala — Updated the Scala-side CometDataWritingCommand to detect S3/S3A output paths and route them through the new native write code path instead of delegating to the JVM Spark writer. Passes the necessary S3 credentials and configuration from Hadoop/Spark config to the native layer.
How are these changes tested?
local testing