ENH: Allow clip, clip_lower, and clip_upper to use array-like thresholds (GH 6966)

Author: Evan Wright

doc/source/whatsnew/v0.16.1.txt

@@ -23,6 +23,7 @@ Enhancements
 
 - Added ``StringMethods.capitalize()`` and ``swapcase`` which behave as the same as standard ``str`` (:issue:`9766`)
 - Added ``StringMethods`` (.str accessor) to ``Index`` (:issue:`9068`)
+- Allow clip, clip_lower, and clip_upper to accept array-like arguments as thresholds (:issue:`6966`). These methods now have an ``axis`` parameter which determines how the Series or DataFrame will be aligned with the threshold(s).
 
   The ``.str`` accessor is now available for both ``Series`` and ``Index``.
 

pandas/core/generic.py

@@ -2821,37 +2821,77 @@ def notnull(self):
         """
         return notnull(self).__finalize__(self)
 
-    def clip(self, lower=None, upper=None, out=None):
+    def clip(self, lower=None, upper=None, out=None, axis=None):
         """
         Trim values at input threshold(s)
 
         Parameters
         ----------
-        lower : float, default None
-        upper : float, default None
+        lower : float or array_like, default None
+        upper : float or array_like, default None
+        axis : int or string axis name, optional
+            Align object with lower and upper along the given axis.
 
         Returns
         -------
         clipped : Series
+
+        Examples
+        --------
+        >>> df
+          0         1
+        0  0.335232 -1.256177
+        1 -1.367855  0.746646
+        2  0.027753 -1.176076
+        3  0.230930 -0.679613
+        4  1.261967  0.570967
+        >>> df.clip(-1.0, 0.5)
+                  0         1
+        0  0.335232 -1.000000
+        1 -1.000000  0.500000
+        2  0.027753 -1.000000
+        3  0.230930 -0.679613
+        4  0.500000  0.500000
+        >>> t
+        0   -0.3
+        1   -0.2
+        2   -0.1
+        3    0.0
+        4    0.1
+        dtype: float64
+        >>> df.clip(t, t + 1, axis=0)
+                  0         1
+        0  0.335232 -0.300000
+        1 -0.200000  0.746646
+        2  0.027753 -0.100000
+        3  0.230930  0.000000
+        4  1.100000  0.570967
         """
         if out is not None:  # pragma: no cover
             raise Exception('out argument is not supported yet')
 
         # GH 2747 (arguments were reversed)
         if lower is not None and upper is not None:
-            lower, upper = min(lower, upper), max(lower, upper)
+            if lib.isscalar(lower) and lib.isscalar(upper):
+                lower, upper = min(lower, upper), max(lower, upper)
 
         result = self
         if lower is not None:
-            result = result.clip_lower(lower)
+            result = result.clip_lower(lower, axis)
         if upper is not None:
-            result = result.clip_upper(upper)
+            result = result.clip_upper(upper, axis)
 
         return result
 
-    def clip_upper(self, threshold):
+    def clip_upper(self, threshold, axis=None):
         """
-        Return copy of input with values above given value truncated
+        Return copy of input with values above given value(s) truncated
+
+        Parameters
+        ----------
+        threshold : float or array_like
+        axis : int or string axis name, optional
+            Align object with threshold along the given axis.
 
         See also
         --------
@@ -2861,14 +2901,21 @@ def clip_upper(self, threshold):
         -------
         clipped : same type as input
         """
-        if isnull(threshold):
+        if np.any(isnull(threshold)):
             raise ValueError("Cannot use an NA value as a clip threshold")
 
-        return self.where((self <= threshold) | isnull(self), threshold)
+        subset = self.le(threshold, axis=axis) | isnull(self)
+        return self.where(subset, threshold, axis=axis)
 
-    def clip_lower(self, threshold):
+    def clip_lower(self, threshold, axis=None):
         """
-        Return copy of the input with values below given value truncated
+        Return copy of the input with values below given value(s) truncated
+
+        Parameters
+        ----------
+        threshold : float or array_like
+        axis : int or string axis name, optional
+            Align object with threshold along the given axis.
 
         See also
         --------
@@ -2878,10 +2925,11 @@ def clip_lower(self, threshold):
         -------
         clipped : same type as input
         """
-        if isnull(threshold):
+        if np.any(isnull(threshold)):
             raise ValueError("Cannot use an NA value as a clip threshold")
 
-        return self.where((self >= threshold) | isnull(self), threshold)
+        subset = self.ge(threshold, axis=axis) | isnull(self)
+        return self.where(subset, threshold, axis=axis)
 
     def groupby(self, by=None, axis=0, level=None, as_index=True, sort=True,
                 group_keys=True, squeeze=False):

pandas/core/ops.py

@@ -571,7 +571,11 @@ def na_op(x, y):
 
         return result
 
-    def wrapper(self, other):
+    def wrapper(self, other, axis=None):
+        # Validate the axis parameter
+        if axis is not None:
+            self._get_axis_number(axis)
+
         if isinstance(other, pd.Series):
             name = _maybe_match_name(self, other)
             if len(self) != len(other):

pandas/tests/test_frame.py

@@ -11394,6 +11394,39 @@ def test_dataframe_clip(self):
             self.assertTrue((clipped_df.values[ub_mask] == ub).all() == True)
             self.assertTrue((clipped_df.values[mask] == df.values[mask]).all() == True)
 
+    def test_clip_against_series(self):
+        # GH #6966
+
+        df = DataFrame(np.random.randn(1000, 2))
+        lb = Series(np.random.randn(1000))
+        ub = lb + 1
+
+        clipped_df = df.clip(lb, ub, axis=0)
+
+        for i in range(2):
+            lb_mask = df.iloc[:, i] <= lb
+            ub_mask = df.iloc[:, i] >= ub
+            mask = ~lb_mask & ~ub_mask
+
+            assert_series_equal(clipped_df.loc[lb_mask, i], lb[lb_mask])
+            assert_series_equal(clipped_df.loc[ub_mask, i], ub[ub_mask])
+            assert_series_equal(clipped_df.loc[mask, i], df.loc[mask, i])
+
+    def test_clip_against_frame(self):
+        df = DataFrame(np.random.randn(1000, 2))
+        lb = DataFrame(np.random.randn(1000, 2))
+        ub = lb + 1
+
+        clipped_df = df.clip(lb, ub)
+
+        lb_mask = df <= lb
+        ub_mask = df >= ub
+        mask = ~lb_mask & ~ub_mask
+
+        assert_frame_equal(clipped_df[lb_mask], lb[lb_mask])
+        assert_frame_equal(clipped_df[ub_mask], ub[ub_mask])
+        assert_frame_equal(clipped_df[mask], df[mask])
+
     def test_get_X_columns(self):
         # numeric and object columns
 

pandas/tests/test_series.py

@@ -5037,6 +5037,20 @@ def test_clip_types_and_nulls(self):
             self.assertEqual(list(isnull(s)), list(isnull(l)))
             self.assertEqual(list(isnull(s)), list(isnull(u)))
 
+    def test_clip_against_series(self):
+        # GH #6966
+
+        s = Series([1.0, 1.0, 4.0])
+        threshold = Series([1.0, 2.0, 3.0])
+
+        assert_series_equal(s.clip_lower(threshold), Series([1.0, 2.0, 4.0]))
+        assert_series_equal(s.clip_upper(threshold), Series([1.0, 1.0, 3.0]))
+
+        lower = Series([1.0, 2.0, 3.0])
+        upper = Series([1.5, 2.5, 3.5])
+        assert_series_equal(s.clip(lower, upper), Series([1.0, 2.0, 3.5]))
+        assert_series_equal(s.clip(1.5, upper), Series([1.5, 1.5, 3.5]))
+
     def test_valid(self):
         ts = self.ts.copy()
         ts[::2] = np.NaN