BUG: Ensure df.itertuples() uses plain tuples correctly

doc/source/whatsnew/v1.0.0.rst

@@ -987,6 +987,7 @@ Other
 - Bug in :class:`Index` where a non-hashable name could be set without raising ``TypeError`` (:issue:`29069`)
 - Bug in :class:`DataFrame` constructor when passing a 2D ``ndarray`` and an extension dtype (:issue:`12513`)
 - Bug in :meth:`DaataFrame.to_csv` when supplied a series with a ``dtype="string"`` and a ``na_rep``, the ``na_rep`` was being truncated to 2 characters. (:issue:`29975`)
+- Bug where :meth:`DataFrame.itertuples` would incorrectly determine whether or not namedtuples could be used for dataframes of 255 columns (:issue:`28282`)
 
 .. _whatsnew_1000.contributors:
 

pandas/core/frame.py

@@ -38,6 +38,7 @@
 
 from pandas._libs import algos as libalgos, lib
 from pandas._typing import Axes, Dtype, FilePathOrBuffer
+from pandas.compat import PY37
 from pandas.compat._optional import import_optional_dependency
 from pandas.compat.numpy import function as nv
 from pandas.util._decorators import (
@@ -975,7 +976,8 @@ def itertuples(self, index=True, name="Pandas"):
         -----
         The column names will be renamed to positional names if they are
         invalid Python identifiers, repeated, or start with an underscore.
-        With a large number of columns (>255), regular tuples are returned.
+        On python versions < 3.7 regular tuples are returned for DataFrames
+        with a large number of columns (>254).
 
         Examples
         --------
@@ -1018,8 +1020,9 @@ def itertuples(self, index=True, name="Pandas"):
         # use integer indexing because of possible duplicate column names
         arrays.extend(self.iloc[:, k] for k in range(len(self.columns)))
 
-        # Python 3 supports at most 255 arguments to constructor
-        if name is not None and len(self.columns) + index < 256:
+        # Python versions before 3.7 support at most 255 arguments to constructors
+        can_return_named_tuples = PY37 or len(self.columns) + index < 255
+        if name is not None and can_return_named_tuples:
             itertuple = collections.namedtuple(name, fields, rename=True)
             return map(itertuple._make, zip(*arrays))
 

pandas/tests/frame/test_api.py

@@ -5,6 +5,8 @@
 import numpy as np
 import pytest
 
+from pandas.compat import PY37
+
 import pandas as pd
 from pandas import Categorical, DataFrame, Series, compat, date_range, timedelta_range
 import pandas.util.testing as tm
@@ -261,8 +263,27 @@ def test_itertuples(self, float_frame):
         df3 = DataFrame({"f" + str(i): [i] for i in range(1024)})
         # will raise SyntaxError if trying to create namedtuple
         tup3 = next(df3.itertuples())
-        assert not hasattr(tup3, "_fields")
         assert isinstance(tup3, tuple)
+        if PY37:
+            assert hasattr(tup3, "_fields")
+        else:
+            assert not hasattr(tup3, "_fields")
+
+        # GH 28282
+        df_254_columns = DataFrame([{f"foo_{i}": f"bar_{i}" for i in range(254)}])
+        result_254_columns = next(df_254_columns.itertuples(index=False))
+        assert isinstance(result_254_columns, tuple)
+        assert hasattr(result_254_columns, "_fields")
+
+        df_255_columns = DataFrame([{f"foo_{i}": f"bar_{i}" for i in range(255)}])
+        result_255_columns = next(df_255_columns.itertuples(index=False))
+        assert isinstance(result_255_columns, tuple)
+
+        # Dataframes with >=255 columns will fallback to regular tuples on python < 3.7
+        if PY37:
+            assert hasattr(result_255_columns, "_fields")
+        else:
+            assert not hasattr(result_255_columns, "_fields")
 
     def test_sequence_like_with_categorical(self):