ENH: df.assign accepting dependent **kwargs (pandas-dev#14207)

Specifically, 'df.assign(b=1, c=lambda x:x['b'])'
does not throw an exception in python 3.6 and above.
Further details are discussed in Issues pandas-dev#14207 and pandas-dev#18797.

Author: datajanko

doc/source/whatsnew/v0.22.0.txt

@@ -139,6 +139,7 @@ Other Enhancements
 - :func:`read_excel()` has gained the ``nrows`` parameter (:issue:`16645`)
 - :func:``DataFrame.to_json`` and ``Series.to_json`` now accept an ``index`` argument which allows the user to exclude the index from the JSON output (:issue:`17394`)
 - ``IntervalIndex.to_tuples()`` has gained the ``na_tuple`` parameter to control whether NA is returned as a tuple of NA, or NA itself (:issue:`18756`)
+- :func:``DataFrame.assign()`` now acceepts dependent kwargs, e.g. `df.assign(b=1, c=lambda x:x['b'])` does not throw an exception anymore. (:issue: `14207)
 
 .. _whatsnew_0220.api_breaking:
 

pandas/core/frame.py

@@ -2659,8 +2659,11 @@ def assign(self, **kwargs):
         \*\*kwargs. For python 3.5 and earlier, since \*\*kwargs is unordered,
         the columns are inserted in alphabetical order at the end of your
         DataFrame.  Assigning multiple columns within the same ``assign``
-        is possible, but you cannot reference other columns created within
-        the same ``assign`` call.
+        is possible, but for python 3.5 and eralier you cannot reference other
+        columns created within the same ``assign`` call. For python 3.6 and
+        above it is possible to reference columns created in an assignment.
+        To this end you have to respect the order of |*|*kwargs and use
+        callables referencing the assigned columns.
 
         Examples
         --------
@@ -2699,19 +2702,21 @@ def assign(self, **kwargs):
         """
         data = self.copy()
 
-        # do all calculations first...
-        results = OrderedDict()
-        for k, v in kwargs.items():
-            results[k] = com._apply_if_callable(v, data)
-
-        # preserve order for 3.6 and later, but sort by key for 3.5 and earlier
+        # for 3.6 preserve order of kwargs
         if PY36:
-            results = results.items()
+            for k, v in kwargs.items():
+                data[k] = com._apply_if_callable(v, data)
         else:
-            results = sorted(results.items())
-        # ... and then assign
-        for k, v in results:
-            data[k] = v
+            # for 3.5 or earlier: do all calculations first...
+            results = OrderedDict()
+            for k, v in kwargs.items():
+                results[k] = com._apply_if_callable(v, data)
+
+            # sort by key for 3.5 and earlier
+                results = sorted(results.items())
+            # ... and then assign
+            for k, v in results:
+                data[k] = v
         return data
 
     def _sanitize_column(self, key, value, broadcast=True):

pandas/tests/frame/test_mutate_columns.py

@@ -89,10 +89,26 @@ def test_assign_bad(self):
             df.assign(lambda x: x.A)
         with pytest.raises(AttributeError):
             df.assign(C=df.A, D=df.A + df.C)
-        with pytest.raises(KeyError):
-            df.assign(C=lambda df: df.A, D=lambda df: df['A'] + df['C'])
-        with pytest.raises(KeyError):
-            df.assign(C=df.A, D=lambda x: x['A'] + x['C'])
+        if not PY36:
+            with pytest.raises(KeyError):
+                df.assign(C=lambda df: df.A,
+                          D=lambda df: df['A'] + df['C'])
+            with pytest.raises(KeyError):
+                df.assign(C=df.A, D=lambda x: x['A'] + x['C'])
+
+    def test_assign_dependent(self):
+        df = DataFrame({'A': [1, 2], 'B': [3, 4]})
+        if PY36:
+            result = df.assign(C=df.A, D=lambda x: x['A'] + x['C'])
+            expected = DataFrame([[1, 3, 1, 2], [2, 4, 2, 4]],
+                                 columns=list('ABCD'))
+            assert_frame_equal(result, expected)
+
+            result = df.assign(C=lambda df: df.A,
+                               D=lambda df: df['A'] + df['C'])
+            expected = DataFrame([[1, 3, 1, 2], [2, 4, 2, 4]],
+                                 columns=list('ABCD'))
+            assert_frame_equal(result, expected)
 
     def test_insert_error_msmgs(self):