Backport PR pandas-dev#56616: BUG: Add limit_area to EA ffill/bfill

Author: rhshadrach

doc/source/whatsnew/v2.2.0.rst

@@ -321,7 +321,7 @@ Other enhancements
 - :meth:`DataFrame.apply` now allows the usage of numba (via ``engine="numba"``) to JIT compile the passed function, allowing for potential speedups (:issue:`54666`)
 - :meth:`ExtensionArray._explode` interface method added to allow extension type implementations of the ``explode`` method (:issue:`54833`)
 - :meth:`ExtensionArray.duplicated` added to allow extension type implementations of the ``duplicated`` method (:issue:`55255`)
-- :meth:`Series.ffill`, :meth:`Series.bfill`, :meth:`DataFrame.ffill`, and :meth:`DataFrame.bfill` have gained the argument ``limit_area`` (:issue:`56492`)
+- :meth:`Series.ffill`, :meth:`Series.bfill`, :meth:`DataFrame.ffill`, and :meth:`DataFrame.bfill` have gained the argument ``limit_area``; 3rd party :class:`.ExtensionArray` authors need to add this argument to the method ``_pad_or_backfill`` (:issue:`56492`)
 - Allow passing ``read_only``, ``data_only`` and ``keep_links`` arguments to openpyxl using ``engine_kwargs`` of :func:`read_excel` (:issue:`55027`)
 - Implement masked algorithms for :meth:`Series.value_counts` (:issue:`54984`)
 - Implemented :meth:`Series.dt` methods and attributes for :class:`ArrowDtype` with ``pyarrow.duration`` type (:issue:`52284`)

pandas/core/arrays/_mixins.py

@@ -305,7 +305,12 @@ def _fill_mask_inplace(
         func(self._ndarray.T, limit=limit, mask=mask.T)
 
     def _pad_or_backfill(
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         mask = self.isna()
         if mask.any():
@@ -315,7 +320,7 @@ def _pad_or_backfill(
             npvalues = self._ndarray.T
             if copy:
                 npvalues = npvalues.copy()
-            func(npvalues, limit=limit, mask=mask.T)
+            func(npvalues, limit=limit, limit_area=limit_area, mask=mask.T)
             npvalues = npvalues.T
 
             if copy:

pandas/core/arrays/arrow/array.py

@@ -1005,13 +1005,18 @@ def dropna(self) -> Self:
         return type(self)(pc.drop_null(self._pa_array))
 
     def _pad_or_backfill(
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         if not self._hasna:
             # TODO(CoW): Not necessary anymore when CoW is the default
             return self.copy()
 
-        if limit is None:
+        if limit is None and limit_area is None:
             method = missing.clean_fill_method(method)
             try:
                 if method == "pad":
@@ -1027,7 +1032,9 @@ def _pad_or_backfill(
 
         # TODO(3.0): after EA.fillna 'method' deprecation is enforced, we can remove
         #  this method entirely.
-        return super()._pad_or_backfill(method=method, limit=limit, copy=copy)
+        return super()._pad_or_backfill(
+            method=method, limit=limit, limit_area=limit_area, copy=copy
+        )
 
     @doc(ExtensionArray.fillna)
     def fillna(

pandas/core/arrays/base.py

@@ -70,6 +70,7 @@
     unique,
 )
 from pandas.core.array_algos.quantile import quantile_with_mask
+from pandas.core.missing import _fill_limit_area_1d
 from pandas.core.sorting import (
     nargminmax,
     nargsort,
@@ -954,7 +955,12 @@ def interpolate(
         )
 
     def _pad_or_backfill(
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         """
         Pad or backfill values, used by Series/DataFrame ffill and bfill.
@@ -1012,6 +1018,12 @@ def _pad_or_backfill(
                 DeprecationWarning,
                 stacklevel=find_stack_level(),
             )
+            if limit_area is not None:
+                raise NotImplementedError(
+                    f"{type(self).__name__} does not implement limit_area "
+                    "(added in pandas 2.2). 3rd-party ExtnsionArray authors "
+                    "need to add this argument to _pad_or_backfill."
+                )
             return self.fillna(method=method, limit=limit)
 
         mask = self.isna()
@@ -1021,6 +1033,8 @@ def _pad_or_backfill(
             meth = missing.clean_fill_method(method)
 
             npmask = np.asarray(mask)
+            if limit_area is not None and not npmask.all():
+                _fill_limit_area_1d(npmask, limit_area)
             if meth == "pad":
                 indexer = libalgos.get_fill_indexer(npmask, limit=limit)
                 return self.take(indexer, allow_fill=True)

pandas/core/arrays/interval.py

@@ -890,11 +890,18 @@ def max(self, *, axis: AxisInt | None = None, skipna: bool = True) -> IntervalOr
         return obj[indexer]
 
     def _pad_or_backfill(  # pylint: disable=useless-parent-delegation
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         # TODO(3.0): after EA.fillna 'method' deprecation is enforced, we can remove
         #  this method entirely.
-        return super()._pad_or_backfill(method=method, limit=limit, copy=copy)
+        return super()._pad_or_backfill(
+            method=method, limit=limit, limit_area=limit_area, copy=copy
+        )
 
     def fillna(
         self, value=None, method=None, limit: int | None = None, copy: bool = True

pandas/core/arrays/masked.py

@@ -192,7 +192,12 @@ def __getitem__(self, item: PositionalIndexer) -> Self | Any:
         return self._simple_new(self._data[item], newmask)
 
     def _pad_or_backfill(
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         mask = self._mask
 
@@ -204,7 +209,21 @@ def _pad_or_backfill(
             if copy:
                 npvalues = npvalues.copy()
                 new_mask = new_mask.copy()
+            elif limit_area is not None:
+                mask = mask.copy()
             func(npvalues, limit=limit, mask=new_mask)
+
+            if limit_area is not None and not mask.all():
+                mask = mask.T
+                neg_mask = ~mask
+                first = neg_mask.argmax()
+                last = len(neg_mask) - neg_mask[::-1].argmax() - 1
+                if limit_area == "inside":
+                    new_mask[:first] |= mask[:first]
+                    new_mask[last + 1 :] |= mask[last + 1 :]
+                elif limit_area == "outside":
+                    new_mask[first + 1 : last] |= mask[first + 1 : last]
+
             if copy:
                 return self._simple_new(npvalues.T, new_mask.T)
             else:

pandas/core/arrays/period.py

@@ -810,12 +810,19 @@ def searchsorted(
         return m8arr.searchsorted(npvalue, side=side, sorter=sorter)
 
     def _pad_or_backfill(
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         # view as dt64 so we get treated as timelike in core.missing,
         #  similar to dtl._period_dispatch
         dta = self.view("M8[ns]")
-        result = dta._pad_or_backfill(method=method, limit=limit, copy=copy)
+        result = dta._pad_or_backfill(
+            method=method, limit=limit, limit_area=limit_area, copy=copy
+        )
         if copy:
             return cast("Self", result.view(self.dtype))
         else:

pandas/core/arrays/sparse/array.py

@@ -716,11 +716,18 @@ def isna(self) -> Self:  # type: ignore[override]
         return type(self)(mask, fill_value=False, dtype=dtype)
 
     def _pad_or_backfill(  # pylint: disable=useless-parent-delegation
-        self, *, method: FillnaOptions, limit: int | None = None, copy: bool = True
+        self,
+        *,
+        method: FillnaOptions,
+        limit: int | None = None,
+        limit_area: Literal["inside", "outside"] | None = None,
+        copy: bool = True,
     ) -> Self:
         # TODO(3.0): We can remove this method once deprecation for fillna method
         #  keyword is enforced.
-        return super()._pad_or_backfill(method=method, limit=limit, copy=copy)
+        return super()._pad_or_backfill(
+            method=method, limit=limit, limit_area=limit_area, copy=copy
+        )
 
     def fillna(
         self,

pandas/core/internals/blocks.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 from functools import wraps
+import inspect
 import re
 from typing import (
     TYPE_CHECKING,
@@ -2256,11 +2257,21 @@ def pad_or_backfill(
     ) -> list[Block]:
         values = self.values
 
+        kwargs: dict[str, Any] = {"method": method, "limit": limit}
+        if "limit_area" in inspect.signature(values._pad_or_backfill).parameters:
+            kwargs["limit_area"] = limit_area
+        elif limit_area is not None:
+            raise NotImplementedError(
+                f"{type(values).__name__} does not implement limit_area "
+                "(added in pandas 2.2). 3rd-party ExtnsionArray authors "
+                "need to add this argument to _pad_or_backfill."
+            )
+
         if values.ndim == 2 and axis == 1:
             # NDArrayBackedExtensionArray.fillna assumes axis=0
-            new_values = values.T._pad_or_backfill(method=method, limit=limit).T
+            new_values = values.T._pad_or_backfill(**kwargs).T
         else:
-            new_values = values._pad_or_backfill(method=method, limit=limit)
+            new_values = values._pad_or_backfill(**kwargs)
         return [self.make_block_same_class(new_values)]