Merge branch 'main' into weekday-fix

Author: mcgeestocks

doc/source/whatsnew/v2.1.0.rst

@@ -493,6 +493,7 @@ Reshaping
 - Bug in :func:`merge_asof` raising ``KeyError`` for extension dtypes (:issue:`52904`)
 - Bug in :func:`merge_asof` raising ``ValueError`` for data backed by read-only ndarrays (:issue:`53513`)
 - Bug in :meth:`DataFrame.agg` and :meth:`Series.agg` on non-unique columns would return incorrect type when dist-like argument passed in (:issue:`51099`)
+- Bug in :meth:`DataFrame.combine_first` ignoring other's columns if ``other`` is empty (:issue:`53792`)
 - Bug in :meth:`DataFrame.idxmin` and :meth:`DataFrame.idxmax`, where the axis dtype would be lost for empty frames (:issue:`53265`)
 - Bug in :meth:`DataFrame.merge` not merging correctly when having ``MultiIndex`` with single level (:issue:`52331`)
 - Bug in :meth:`DataFrame.stack` losing extension dtypes when columns is a :class:`MultiIndex` and frame contains mixed dtypes (:issue:`45740`)

pandas/_libs/groupby.pyx

@@ -950,7 +950,7 @@ def group_skew(
                     isna_entry = _treat_as_na(val, False)
 
                 if not isna_entry:
-                    # Based on RunningSats::Push from
+                    # Based on RunningStats::Push from
                     #  https://www.johndcook.com/blog/skewness_kurtosis/
                     n1 = nobs[lab, j]
                     n = n1 + 1

pandas/_libs/tslibs/timedeltas.pyx

@@ -974,7 +974,7 @@ cdef _timedelta_from_value_and_reso(cls, int64_t value, NPY_DATETIMEUNIT reso):
             "Only resolutions 's', 'ms', 'us', 'ns' are supported."
         )
 
-    td_base._value= value
+    td_base._value = value
     td_base._is_populated = 0
     td_base._creso = reso
     return td_base

pandas/core/arrays/arrow/array.py

@@ -1131,13 +1131,7 @@ def take(
         it's called by :meth:`Series.reindex`, or any other method
         that causes realignment, with a `fill_value`.
         """
-        # TODO: Remove once we got rid of the (indices < 0) check
-        if not is_array_like(indices):
-            indices_array = np.asanyarray(indices)
-        else:
-            # error: Incompatible types in assignment (expression has type
-            # "Sequence[int]", variable has type "ndarray")
-            indices_array = indices  # type: ignore[assignment]
+        indices_array = np.asanyarray(indices)
 
         if len(self._pa_array) == 0 and (indices_array >= 0).any():
             raise IndexError("cannot do a non-empty take")

pandas/core/arrays/sparse/array.py

@@ -42,7 +42,6 @@
     maybe_box_datetimelike,
 )
 from pandas.core.dtypes.common import (
-    is_array_like,
     is_bool_dtype,
     is_integer,
     is_list_like,
@@ -428,19 +427,16 @@ def __init__(
             # Union[int, Sequence[int]]], List[Any], _DTypeDict, Tuple[Any, Any]]]"
             data = np.array([], dtype=dtype)  # type: ignore[arg-type]
 
-        if not is_array_like(data):
-            try:
-                # probably shared code in sanitize_series
-
-                data = sanitize_array(data, index=None)
-            except ValueError:
-                # NumPy may raise a ValueError on data like [1, []]
-                # we retry with object dtype here.
-                if dtype is None:
-                    dtype = np.dtype(object)
-                    data = np.atleast_1d(np.asarray(data, dtype=dtype))
-                else:
-                    raise
+        try:
+            data = sanitize_array(data, index=None)
+        except ValueError:
+            # NumPy may raise a ValueError on data like [1, []]
+            # we retry with object dtype here.
+            if dtype is None:
+                dtype = np.dtype(object)
+                data = np.atleast_1d(np.asarray(data, dtype=dtype))
+            else:
+                raise
 
         if copy:
             # TODO: avoid double copy when dtype forces cast.

pandas/core/dtypes/dtypes.py

@@ -1665,17 +1665,6 @@ def _check_fill_value(self):
                     FutureWarning,
                     stacklevel=find_stack_level(),
                 )
-        elif isinstance(self.subtype, CategoricalDtype):
-            # TODO: is this even supported?  It is reached in
-            #  test_dtype_sparse_with_fill_value_not_present_in_data
-            if self.subtype.categories is None or val not in self.subtype.categories:
-                warnings.warn(
-                    "Allowing arbitrary scalar fill_value in SparseDtype is "
-                    "deprecated. In a future version, the fill_value must be "
-                    "a valid value for the SparseDtype.subtype.",
-                    FutureWarning,
-                    stacklevel=find_stack_level(),
-                )
         else:
             dummy = np.empty(0, dtype=self.subtype)
             dummy = ensure_wrapped_if_datetimelike(dummy)

pandas/core/frame.py

@@ -673,9 +673,9 @@ def __init__(
         manager = get_option("mode.data_manager")
 
         # GH47215
-        if index is not None and isinstance(index, set):
+        if isinstance(index, set):
             raise ValueError("index cannot be a set")
-        if columns is not None and isinstance(columns, set):
+        if isinstance(columns, set):
             raise ValueError("columns cannot be a set")
 
         if copy is None:
@@ -8344,7 +8344,13 @@ def combiner(x, y):
 
             return expressions.where(mask, y_values, x_values)
 
-        combined = self.combine(other, combiner, overwrite=False)
+        if len(other) == 0:
+            combined = self.reindex(
+                self.columns.append(other.columns.difference(self.columns)), axis=1
+            )
+            combined = combined.astype(other.dtypes)
+        else:
+            combined = self.combine(other, combiner, overwrite=False)
 
         dtypes = {
             col: find_common_type([self.dtypes[col], other.dtypes[col]])

pandas/core/generic.py

@@ -47,6 +47,7 @@
     AlignJoin,
     AnyArrayLike,
     ArrayLike,
+    Axes,
     Axis,
     AxisInt,
     CompressionOptions,
@@ -271,7 +272,7 @@ def __init__(self, data: Manager) -> None:
     def _init_mgr(
         cls,
         mgr: Manager,
-        axes,
+        axes: dict[Literal["index", "columns"], Axes | None],
         dtype: DtypeObj | None = None,
         copy: bool_t = False,
     ) -> Manager:
@@ -3995,7 +3996,6 @@ class  max_speed
             ):
                 return self.copy(deep=None)
         elif self.ndim == 1:
-            # TODO: be consistent here for DataFrame vs Series
             raise TypeError(
                 f"{type(self).__name__}.take requires a sequence of integers, "
                 "not slice."

pandas/core/indexes/base.py

@@ -4192,7 +4192,7 @@ def _convert_slice_indexer(self, key: slice, kind: Literal["loc", "getitem"]):
 
         # TODO(GH#50617): once Series.__[gs]etitem__ is removed we should be able
         #  to simplify this.
-        if isinstance(self.dtype, np.dtype) and self.dtype.kind == "f":
+        if lib.is_np_dtype(self.dtype, "f"):
             # We always treat __getitem__ slicing as label-based
             # translate to locations
             return self.slice_indexer(start, stop, step)

pandas/core/internals/blocks.py

@@ -233,7 +233,7 @@ def make_block_same_class(
         values,
         placement: BlockPlacement | None = None,
         refs: BlockValuesRefs | None = None,
-    ) -> Block:
+    ) -> Self:
         """Wrap given values in a block of same type as self."""
         # Pre-2.0 we called ensure_wrapped_if_datetimelike because fastparquet
         #  relied on it, as of 2.0 the caller is responsible for this.
@@ -503,6 +503,7 @@ def convert(
     # ---------------------------------------------------------------------
     # Array-Like Methods
 
+    @final
     @cache_readonly
     def dtype(self) -> DtypeObj:
         return self.values.dtype
@@ -559,7 +560,7 @@ def to_native_types(self, na_rep: str = "nan", quoting=None, **kwargs) -> Block:
         return self.make_block(result)
 
     @final
-    def copy(self, deep: bool = True) -> Block:
+    def copy(self, deep: bool = True) -> Self:
         """copy constructor"""
         values = self.values
         refs: BlockValuesRefs | None
@@ -1499,7 +1500,8 @@ def quantile(
         result = ensure_block_shape(result, ndim=2)
         return new_block_2d(result, placement=self._mgr_locs)
 
-    def round(self, decimals: int, using_cow: bool = False) -> Block:
+    @final
+    def round(self, decimals: int, using_cow: bool = False) -> Self:
         """
         Rounds the values.
         If the block is not of an integer or float dtype, nothing happens.
@@ -1765,9 +1767,7 @@ def putmask(self, mask, new, using_cow: bool = False) -> list[Block]:
 
         if using_cow and self.refs.has_reference():
             values = values.copy()
-            self = self.make_block_same_class(  # type: ignore[assignment]
-                values.T if values.ndim == 2 else values
-            )
+            self = self.make_block_same_class(values.T if values.ndim == 2 else values)
 
         try:
             # Caller is responsible for ensuring matching lengths

pandas/core/internals/concat.py

@@ -254,6 +254,15 @@ def _concat_homogeneous_fastpath(
     """
     # assumes
     #  all(_is_homogeneous_mgr(mgr, first_dtype) for mgr, _ in in mgrs_indexers)
+
+    if all(not indexers for _, indexers in mgrs_indexers):
+        # https://github.com/pandas-dev/pandas/pull/52685#issuecomment-1523287739
+        arrs = [mgr.blocks[0].values.T for mgr, _ in mgrs_indexers]
+        arr = np.concatenate(arrs).T
+        bp = libinternals.BlockPlacement(slice(shape[0]))
+        nb = new_block_2d(arr, bp)
+        return nb
+
     arr = np.empty(shape, dtype=first_dtype)
 
     if first_dtype == np.float64:

pandas/core/internals/managers.py

@@ -939,8 +939,7 @@ def take(
         -------
         BlockManager
         """
-        assert isinstance(indexer, np.ndarray), type(indexer)
-        assert indexer.dtype == np.intp, indexer.dtype
+        # Caller is responsible for ensuring indexer annotation is accurate
 
         n = self.shape[axis]
         indexer = maybe_convert_indices(indexer, n, verify=verify)

pandas/core/missing.py

@@ -311,8 +311,6 @@ def interpolate_array_2d(
     limit_direction: str = "forward",
     limit_area: str | None = None,
     fill_value: Any | None = None,
-    coerce: bool = False,
-    downcast: str | None = None,
     **kwargs,
 ) -> None:
     """