Skip to content

[SPARK-58013][PS] Fix loc setitem dropping writes for reordered columns on pandas 3 #57504

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Vivek1106-04 wants to merge 2 commits into
base: master
Choose a base branch
from
+1 −25
Open

[SPARK-58013][PS] Fix loc setitem dropping writes for reordered columns on pandas 3 #57504

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0753c3a
[SPARK-58013][PS] Fix loc setitem dropping writes for reordered colum…
Vivek1106-04 Jul 24, 2026
16bb359
[SPARK-58013][PS] Revert unused unpack name to _ in loc setitem
Vivek1106-04 Jul 25, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 1 addition & 25 deletions python/pyspark/pandas/indexing.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -741,7 +741,7 @@ def __setitem__(self, key: Any, value: Any) -> None:
cond, limit, remaining_index = self._select_rows(rows_sel)
missing_keys: List[Name] = []
(
selected_column_labels,
_,
data_spark_columns,
_,
_,
Expand All @@ -763,30 +763,6 @@ def __setitem__(self, key: Any, value: Any) -> None:
if isinstance(value, Series):
value = value.spark.column
else:
if (
# Only apply this behavior for pandas 3+, where CoW semantics changed.
LooseVersion(pd.__version__) >= "3.0.0"
# Only for multi-column assignment (single-column assignment is unaffected).
and len(selected_column_labels) > 1
# Column selector must be list-like (e.g. ["shield", "max_speed"]), not scalar label access.
and is_list_like(cols_sel)
# Excludes string/bytes (single label), tuple (e.g. MultiIndex label),
# and slice selectors; keeps this narrowly on explicit column lists.
and not isinstance(cols_sel, (str, bytes, tuple, slice))
# Only trigger when cached/anchored Series exist on the frame,
# matching the problematic case where views were materialized before assignment.
and hasattr(self._psdf_or_psser, "_psseries")
):
selected_column_labels_set = set(selected_column_labels)
selected_labels_in_internal_order = [
label
for label in self._internal.column_labels
if label in selected_column_labels_set
]
if selected_column_labels != selected_labels_in_internal_order:
# If requested columns are in different order than the DataFrame's internal order,
# it returns early (no-op), matching pandas 3 behavior for that edge case.
return
value = F.lit(value)

new_data_spark_columns = []
Expand Down