FIX apply check_array before stacking in ColumnTransformer #27671
Conversation
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
glemaitre
reviewed
Oct 30, 2023
Member
|
Could you add an entry to changelog just to make the CI happy :) |
Co-authored-by: Guillaume Lemaitre <[email protected]>
Contributor
Author
|
applying using the current state of this PR: import pandas as pd
from sklearn.compose import make_column_transformer
X = pd.DataFrame({"A": [0.5, 1.3], "B": ['a', 'b']}).convert_dtypes()
transformer = make_column_transformer(("passthrough", ["A", "B"]))
X2 = transformer.fit_transform(X) |
Contributor
Author
|
In the developers meeting it was decided that instead of this approach we should raise an error when transformers' outputs contain pd.NA and output is not configured as "pandas", so I am closing this PR in favor of #27734 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
Reference Issues/PRs
fixes #27482
What does this implement/fix? Explain your changes.
The
ColumnTransformer(withoutset_output(transform="pandas")), when it transforms pandas columns that containpd.NA, outputs arrays of dtypeobjectthat containpd.NA, which can cause subsequent estimators in a pipeline to fail becausecheck_array(dtype="numeric")raises a TypeError on such arrays.The proposal here is to apply
check_array(dtype=None)on the transformers' outputs before performing the horizontal stacking. This will convertpd.Float64tonp.float64andpd.NAtonp.nanAny other comments?
Another option is to forbid
pd.NAin the individual transformers' output if the ColumnTransformer does not have its output transform set to "pandas". Before performing thehstackwe can check forpd.NAand raise if any are found.This still has the drawback that the output for
pd.Float64columns (without missing values) will beobjectrather thannp.float64