ENH: DataFrame.drop copy kwd

doc/source/whatsnew/v1.5.0.rst

@@ -293,6 +293,7 @@ Other enhancements
 - :meth:`RangeIndex.union` now can return a :class:`RangeIndex` instead of a :class:`Int64Index` if the resulting values are equally spaced (:issue:`47557`, :issue:`43885`)
 - :meth:`DataFrame.compare` now accepts an argument ``result_names`` to allow the user to specify the result's names of both left and right DataFrame which are being compared. This is by default ``'self'`` and ``'other'`` (:issue:`44354`)
 - :meth:`Series.add_suffix`, :meth:`DataFrame.add_suffix`, :meth:`Series.add_prefix` and :meth:`DataFrame.add_prefix` support a ``copy`` argument. If ``False``, the underlying data is not copied in the returned object (:issue:`47934`)
+- :meth:`DataFrame.drop` with ``axis=1`` can now accept ``copy=False`` to prevent a copy of the underlying data (:issue:`47993`)
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_150.notable_bug_fixes:

pandas/core/base.py

@@ -221,9 +221,7 @@ def _obj_with_exclusions(self):
         if len(self.exclusions) > 0:
             # equivalent to `self.obj.drop(self.exclusions, axis=1)
             #  but this avoids consolidating and making a copy
-            # TODO: following GH#45287 can we now use .drop directly without
-            #  making a copy?
-            return self.obj._drop_axis(self.exclusions, axis=1, only_slice=True)
+            return self.obj.drop(self.exclusions, axis=1, copy=False)
         else:
             return self.obj
 

pandas/core/frame.py

@@ -5096,6 +5096,7 @@ def drop(
         level: Level | None = ...,
         inplace: Literal[True],
         errors: IgnoreRaise = ...,
+        copy: bool | lib.NoDefault = ...,
     ) -> None:
         ...
 
@@ -5110,6 +5111,7 @@ def drop(
         level: Level | None = ...,
         inplace: Literal[False] = ...,
         errors: IgnoreRaise = ...,
+        copy: bool | lib.NoDefault = ...,
     ) -> DataFrame:
         ...
 
@@ -5124,6 +5126,7 @@ def drop(
         level: Level | None = ...,
         inplace: bool = ...,
         errors: IgnoreRaise = ...,
+        copy: bool | lib.NoDefault = ...,
     ) -> DataFrame | None:
         ...
 
@@ -5139,6 +5142,8 @@ def drop(  # type: ignore[override]
         level: Level | None = None,
         inplace: bool = False,
         errors: IgnoreRaise = "raise",
+        *,
+        copy: bool | lib.NoDefault = lib.no_default,
     ) -> DataFrame | None:
         """
         Drop specified labels from rows or columns.
@@ -5171,6 +5176,10 @@ def drop(  # type: ignore[override]
         errors : {'ignore', 'raise'}, default 'raise'
             If 'ignore', suppress error and only existing labels are
             dropped.
+        copy : bool, default True
+            If False and axis == 1, do not make a copy of the underlying data.
+
+            .. versionadded:: 1.5.0
 
         Returns
         -------
@@ -5285,6 +5294,7 @@ def drop(  # type: ignore[override]
             level=level,
             inplace=inplace,
             errors=errors,
+            copy=copy,
         )
 
     @overload

pandas/core/generic.py

@@ -4376,6 +4376,7 @@ def drop(
         level: Level | None = ...,
         inplace: Literal[True],
         errors: IgnoreRaise = ...,
+        copy: bool_t | lib.NoDefault = ...,
     ) -> None:
         ...
 
@@ -4390,6 +4391,7 @@ def drop(
         level: Level | None = ...,
         inplace: Literal[False] = ...,
         errors: IgnoreRaise = ...,
+        copy: bool_t | lib.NoDefault = ...,
     ) -> NDFrameT:
         ...
 
@@ -4404,6 +4406,7 @@ def drop(
         level: Level | None = ...,
         inplace: bool_t = ...,
         errors: IgnoreRaise = ...,
+        copy: bool_t | lib.NoDefault = ...,
     ) -> NDFrameT | None:
         ...
 
@@ -4417,9 +4420,17 @@ def drop(
         level: Level | None = None,
         inplace: bool_t = False,
         errors: IgnoreRaise = "raise",
+        *,
+        copy: bool_t | lib.NoDefault = lib.no_default,
     ) -> NDFrameT | None:
 
         inplace = validate_bool_kwarg(inplace, "inplace")
+        if inplace:
+            if copy is not lib.no_default:
+                raise ValueError("Cannot pass both inplace=True and copy")
+            copy = True
+        elif copy is lib.no_default:
+            copy = True
 
         if labels is not None:
             if index is not None or columns is not None:
@@ -4437,7 +4448,9 @@ def drop(
 
         for axis, labels in axes.items():
             if labels is not None:
-                obj = obj._drop_axis(labels, axis, level=level, errors=errors)
+                obj = obj._drop_axis(
+                    labels, axis, level=level, errors=errors, only_slice=not copy
+                )
 
         if inplace:
             self._update_inplace(obj)

pandas/core/series.py

@@ -4992,7 +4992,8 @@ def reindex(self, *args, **kwargs) -> Series:
             kwargs.update({"index": index})
         return super().reindex(**kwargs)
 
-    @overload
+    # error: Signature of "drop" incompatible with supertype "NDFrame"  [override]
+    @overload  # type: ignore[override]
     def drop(
         self,
         labels: IndexLabel = ...,
@@ -5141,6 +5142,7 @@ def drop(  # type: ignore[override]
             level=level,
             inplace=inplace,
             errors=errors,
+            copy=lib.no_default,
         )
 
     @overload

pandas/tests/frame/methods/test_drop.py

@@ -67,6 +67,25 @@ def test_drop_with_non_unique_datetime_index_and_invalid_keys():
 
 
 class TestDataFrameDrop:
+    def test_drop_copy(self):
+        df = DataFrame(
+            [[1, 2, 3], [3, 4, 5], [5, 6, 7]],
+            index=["a", "b", "c"],
+            columns=["d", "e", "f"],
+        )
+
+        msg = "Cannot pass both inplace=True and copy"
+        with pytest.raises(ValueError, match=msg):
+            df.drop("d", axis=1, inplace=True, copy=True)
+        with pytest.raises(ValueError, match=msg):
+            df.drop("d", axis=1, inplace=True, copy=False)
+
+        res = df.drop("d", axis=1, copy=True)
+        assert not any(tm.shares_memory(res[c], df[c]) for c in res.columns)
+
+        res = df.drop("d", axis=1, copy=False)
+        assert all(tm.shares_memory(res[c], df[c]) for c in res.columns)
+
     def test_drop_names(self):
         df = DataFrame(
             [[1, 2, 3], [3, 4, 5], [5, 6, 7]],