[Data] Fix to_pandas error when multiple block types
#48583
Merged
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 <bveeramani@berkeley.edu>
Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu>
bveeramani
requested review from
scottjlee,
raulchen,
stephanie-wang,
omatthew98 and
alexeykudinkin
as code owners
scottjlee
reviewed
Nov 6, 2024
Comment on lines
+4617
to
+4625
| builder = PandasBlockBuilder() | ||
| for batch in self.iter_batches(batch_format="pandas", batch_size=None): | ||
| builder.add_block(batch) | ||
| block = builder.build() | ||
|
|
||
| # `PandasBlockBuilder` creates a dataframe with internal extension types like | ||
| # 'TensorDtype'. We use the `to_pandas` method to convert these extension | ||
| # types to regular types. | ||
| return BlockAccessor.for_block(block).to_pandas() |
There was a problem hiding this comment.
if i'm understanding correctly, i think this new code now adds an extra conversion from Arrow to pandas batch. is there any way we can avoid it?
the old code:
Arrow internal -> Arrow Block -> build to big Arrow block -> convert to pd DF
the new code:
Arrow internal -> Pandas batch -> build to big Arrow block -> convert to pd DF
There was a problem hiding this comment.
There's no step 3 ("build to big Arrow block") in the new code, since we're using PandasBlockBuilder
There was a problem hiding this comment.
ah you're right, i mistakenly thought it converts to arrow tables.
alexeykudinkin
approved these changes
Nov 7, 2024
Comment on lines
+4617
to
+4625
| builder = PandasBlockBuilder() | ||
| for batch in self.iter_batches(batch_format="pandas", batch_size=None): | ||
| builder.add_block(batch) | ||
| block = builder.build() | ||
|
|
||
| # `PandasBlockBuilder` creates a dataframe with internal extension types like | ||
| # 'TensorDtype'. We use the `to_pandas` method to convert these extension | ||
| # types to regular types. | ||
| return BlockAccessor.for_block(block).to_pandas() |
There was a problem hiding this comment.
There's no step 3 ("build to big Arrow block") in the new code, since we're using PandasBlockBuilder
scottjlee
approved these changes
Nov 7, 2024
JP-sDEV
pushed a commit
to JP-sDEV/ray
that referenced
this pull request
Nov 14, 2024
…8583) Two issues: 1. When the execution plan caches the dataset schema, it might call `unify_schema` to produce a single schema from all of the bundles' schemas. The issue is that this function expects all of the input schemas to be of Arrow type, but we only check if at least one schema is of Arrow type before calling the function. 2. `to_pandas` iterates over blocks and adds them to `DelegatingBlockBuilder`. The issue is that `DelegatingBlockBuilder` expects all input blocks to be of the same type. --------- Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu>
mohitjain2504
pushed a commit
to mohitjain2504/ray
that referenced
this pull request
Nov 15, 2024
…8583) Two issues: 1. When the execution plan caches the dataset schema, it might call `unify_schema` to produce a single schema from all of the bundles' schemas. The issue is that this function expects all of the input schemas to be of Arrow type, but we only check if at least one schema is of Arrow type before calling the function. 2. `to_pandas` iterates over blocks and adds them to `DelegatingBlockBuilder`. The issue is that `DelegatingBlockBuilder` expects all input blocks to be of the same type. --------- Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu> Signed-off-by: mohitjain2504 <mohit.jain@dream11.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?
Two issues:
unify_schemato produce a single schema from all of the bundles' schemas. The issue is that this function expects all of the input schemas to be of Arrow type, but we only check if at least one schema is of Arrow type before calling the function.to_pandasiterates over blocks and adds them toDelegatingBlockBuilder. The issue is thatDelegatingBlockBuilderexpects all input blocks to be of the same type.Related issue number
Fixes #48575
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.