Merge pull request #9818 from pydata/assign-sort

API: Sort keys for DataFrame.assign

Author: Tom Augspurger

doc/source/dsintro.rst

@@ -461,7 +461,7 @@ Inspired by `dplyr's
 <http://cran.rstudio.com/web/packages/dplyr/vignettes/introduction.html#mutate>`__
 ``mutate`` verb, DataFrame has an :meth:`~pandas.DataFrame.assign`
 method that allows you to easily create new columns that are potentially
-derived from existing columns. 
+derived from existing columns.
 
 .. ipython:: python
 
@@ -511,7 +511,9 @@ DataFrame is returned, with the new values inserted.
 .. warning::
 
   Since the function signature of ``assign`` is ``**kwargs``, a dictionary,
-  the order of the new columns in the resulting DataFrame cannot be guaranteed.
+  the order of the new columns in the resulting DataFrame cannot be guaranteed
+  to match the order you pass in. To make things predictable, items are inserted
+  alphabetically (by key) at the end of the DataFrame.
 
   All expressions are computed first, and then assigned. So you can't refer
   to another column being assigned in the same call to ``assign``. For example:

doc/source/whatsnew/v0.16.1.txt

@@ -45,6 +45,10 @@ API changes
 - Add support for separating years and quarters using dashes, for
   example 2014-Q1.  (:issue:`9688`)
 
+- :meth:`~pandas.DataFrame.assign` now inserts new columns in alphabetical order. Previously
+  the order was arbitrary. (:issue:`9777`)
+
+
 .. _whatsnew_0161.performance:
 
 Performance Improvements

pandas/core/frame.py

@@ -2244,10 +2244,11 @@ def assign(self, **kwargs):
         Notes
         -----
         Since ``kwargs`` is a dictionary, the order of your
-        arguments may not be preserved, and so the order of the
-        new columns is not well defined. Assigning multiple
-        columns within the same ``assign`` is possible, but you cannot
-        reference other columns created within the same ``assign`` call.
+        arguments may not be preserved. The make things predicatable,
+        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.
 
         Examples
         --------
@@ -2296,7 +2297,7 @@ def assign(self, **kwargs):
                 results[k] = v
 
         # ... and then assign
-        for k, v in results.items():
+        for k, v in sorted(results.items()):
             data[k] = v
 
         return data

pandas/tests/test_frame.py

@@ -14073,12 +14073,21 @@ def test_assign(self):
         assert_frame_equal(result, expected)
 
     def test_assign_multiple(self):
-        df = DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
+        df = DataFrame([[1, 4], [2, 5], [3, 6]], columns=['A', 'B'])
         result = df.assign(C=[7, 8, 9], D=df.A, E=lambda x: x.B)
-        expected = DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6], 'C': [7, 8, 9],
-                              'D': [1, 2, 3], 'E': [4, 5, 6]})
-        # column order isn't preserved
-        assert_frame_equal(result.reindex_like(expected), expected)
+        expected = DataFrame([[1, 4, 7, 1, 4], [2, 5, 8, 2, 5],
+                              [3, 6, 9, 3, 6]], columns=list('ABCDE'))
+        assert_frame_equal(result, expected)
+
+    def test_assign_alphabetical(self):
+        # GH 9818
+        df = DataFrame([[1, 2], [3, 4]], columns=['A', 'B'])
+        result = df.assign(D=df.A + df.B, C=df.A - df.B)
+        expected = DataFrame([[1, 2, -1, 3], [3, 4, -1, 7]],
+                             columns=list('ABCD'))
+        assert_frame_equal(result, expected)
+        result = df.assign(C=df.A - df.B, D=df.A + df.B)
+        assert_frame_equal(result, expected)
 
     def test_assign_bad(self):
         df = DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})