Fix Delta Lake atomic table operations on spark341db [databricks] #9729
Conversation
This file contains 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
Signed-off-by: Jason Lowe <[email protected]>
revans2
approved these changes
Nov 15, 2023
|
build |
1 similar comment
|
build |
|
Build failed due to an unrelated test, posted fix for that test at #9745 |
|
build |
1 similar comment
|
build |
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Outdated suggestions cannot be applied.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.
Fixes #9676. Databricks 13.3 includes the changes from SPARK-43088 where atomic operations do not include the query at the point where the staged table is created but instead create then append the data. For Delta Lake, this manifests as a subplan being executed which contains an AppendDataExecV1 using the table and WriteBuilder created to handle the original atomic table operation. That explains why the issue reported seeing a tagging attempt on a RAPIDS Accelerator class.
The fix is relatively straightforward. We go ahead and let the AppendDataExecV1 use the RAPIDS Accelerator classes, but we recognize these classes in the tagging code and skip the CPU extraction code when that is the case. This keeps the append operation on the GPU because we're already using a table and writer geared to the GPU.