[Data] Avoid serializing datasource for Parquet read tasks #41712
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: Balaji Veeramani <balaji@anyscale.com>
bveeramani
requested review from
ericl,
scv119,
c21,
amogkam,
scottjlee,
raulchen,
stephanie-wang and
Zandew
as code owners
|
Is there a way to add a unit test to prevent the issue in the future? |
|
This sounds to me a release blocker. We should cherry pick to prevent regression. |
scottjlee
approved these changes
Dec 8, 2023
raulchen
approved these changes
Dec 8, 2023
raulchen
reviewed
Dec 8, 2023
| @@ -390,7 +391,7 @@ def get_read_tasks(self, parallelism: int) -> List[ReadTask]: | |||
| columns, | |||
| schema, | |||
| f, | |||
| self._include_paths, | |||
| include_paths, | |||
There was a problem hiding this comment.
we can make the ParquetDatasource class non-serializable to avoid this issue in the future.
@c21 I can't think of a good way to directly unit test this. If it's okay with you, I'll merge this PR now since it's a release blocker, and if we come up with an effective test, I'll add it in a follow-up PR? Also, per @raulchen's suggestion, we could make |
bveeramani
added a commit
to bveeramani/ray
that referenced
this pull request
Dec 8, 2023
…ct#41712) ray-project#41118 added an include_paths parameter to ParquetDatasource. As part of the PR, we pass an self._include_paths attribute to Parquet read tasks. As a result, the datasource (self) gets serialized with each read tasks. Normally, this isn't an issue, but if you're working with a large dataset (like in the failing release test), then the datasource is slow to serialize. This PR fixes the issue by removing the reference to self. Signed-off-by: Balaji Veeramani <balaji@anyscale.com>
8 tasks
architkulkarni
pushed a commit
that referenced
this pull request
Dec 8, 2023
…41725) #41118 added an include_paths parameter to ParquetDatasource. As part of the PR, we pass an self._include_paths attribute to Parquet read tasks. As a result, the datasource (self) gets serialized with each read tasks. Normally, this isn't an issue, but if you're working with a large dataset (like in the failing release test), then the datasource is slow to serialize. This PR fixes the issue by removing the reference to self. Signed-off-by: Balaji Veeramani <balaji@anyscale.com>
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.
Why are these changes needed?
#41118 added an
include_pathsparameter toParquetDatasource. As part of the PR, we pass anself._include_pathsattribute to Parquet read tasks:ray/python/ray/data/datasource/parquet_datasource.py
Lines 385 to 394 in e862ec2
As a result, the datasource (
self) gets serialized with each read tasks. Normally, this isn't an issue, but if you're working with a large dataset (like in the failing release test), then the datasource is slow to serialize.This PR fixes the issue by removing the reference to
self.Related issue number
Fixes #41390
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/under thecorresponding
.rstfile.