fix(rust): Preserve List inner dtype during chunked take operations #25634
+42
−5
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
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #25634 +/- ##
==========================================
+ Coverage 79.35% 79.44% +0.09%
==========================================
Files 1743 1743
Lines 240295 240328 +33
Branches 3038 3038
==========================================
+ Hits 190683 190928 +245
+ Misses 48830 48618 -212
Partials 782 782 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| let ca = self.list().unwrap(); | ||
| ca.take_chunked_unchecked(by, sorted, avoid_sharing) | ||
| .into_series() | ||
| let taken = ca.take_chunked_unchecked(by, sorted, avoid_sharing); |
Collaborator
There was a problem hiding this comment.
I don't think this is the right place to fix it - could we instead update take_chunked_unchecked() impl for ChunkedArray to ensure the original type is restored there? -
davidia
force-pushed
the
fix-list-categorical-dtype-loss
branch
3 times, most recently
from
c725f27 to
dc53b56
Compare
When performing left/right joins on chunked DataFrames (common with native Parquet reader), the `take_chunked_unchecked` and `take_opt_chunked_unchecked` methods would lose dtype information for nested types like `List(Categorical)`. The issue was that `ChunkedArray::with_chunk` re-infers the dtype from the physical Arrow array, causing `List(Categorical)` to become `List(UInt32)`. The fix uses `ChunkedArray::with_chunk_like` instead, which preserves the original ChunkedArray's dtype when constructing the result. Fixes pola-rs#25626
davidia
force-pushed
the
fix-list-categorical-dtype-loss
branch
from
dc53b56 to
233d6cc
Compare
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.
When performing left/right joins on chunked DataFrames the
take_chunked_uncheckedandtake_opt_chunked_uncheckedmethods forListtypes would lose the inner dtype information. This causedList(Categorical)to becomeList(UInt32)becauseChunkedArray::with_chunkre-infers the dtype from the physical Arrow array.The fix preserves the original dtype by using
Series::from_chunks_and_dtype_uncheckedwith the originalself.dtype()instead of letting it be re-inferred.Fixes #25626