PERF: speed-up DateFrame.itertuples() with namedtuples

Author: xflr6

asv_bench/benchmarks/frame_methods.py

@@ -653,6 +653,16 @@ def j(self):
             self.df3[0]
 
 
+class frame_itertuples(object):
+
+    def setup(self):
+        self.df = DataFrame(np.random.randn(50000, 10))
+
+    def time_frame_itertuples(self):
+        for row in self.df.itertuples():
+            pass
+
+
 class frame_mask_bools(object):
     goal_time = 0.2
 

doc/source/whatsnew/v0.17.1.txt

@@ -108,7 +108,7 @@ API changes
 - ``Series.sort_index()`` now correctly handles the ``inplace`` option (:issue:`11402`)
 - ``SparseArray.__iter__()`` now does not cause ``PendingDeprecationWarning`` in Python 3.5 (:issue:`11622`)
 
-- ``DataFrame.itertuples()`` now returns ``namedtuple`` objects, when possible. (:issue:`11269`)
+- ``DataFrame.itertuples()`` now returns ``namedtuple`` objects, when possible. (:issue:`11269`, :issue:`11625`)
 - ``Series.ptp`` will now ignore missing values by default (:issue:`11163`)
 
 .. _whatsnew_0171.deprecations:

pandas/core/frame.py

@@ -43,7 +43,7 @@
 import pandas.computation.expressions as expressions
 from pandas.computation.eval import eval as _eval
 from numpy import percentile as _quantile
-from pandas.compat import(range, zip, lrange, lmap, lzip, StringIO, u,
+from pandas.compat import(range, map, zip, lrange, lmap, lzip, StringIO, u,
                           OrderedDict, raise_with_traceback)
 from pandas import compat
 from pandas.sparse.array import SparseArray
@@ -664,7 +664,7 @@ def itertuples(self, index=True, name="Pandas"):
         index : boolean, default True
             If True, return the index as the first element of the tuple.
         name : string, default "Pandas"
-            The name of the returned namedtuple.
+            The name of the returned namedtuples or None to return regular tuples.
 
         Notes
         -----
@@ -703,13 +703,13 @@ def itertuples(self, index=True, name="Pandas"):
 
         # Python 3 supports at most 255 arguments to constructor, and
         # things get slow with this many fields in Python 2
-        if len(self.columns) + index < 256:
+        if name is not None and len(self.columns) + index < 256:
             # `rename` is unsupported in Python 2.6
             try:
                 itertuple = collections.namedtuple(
                     name, fields+list(self.columns), rename=True)
-                return (itertuple(*row) for row in zip(*arrays))
-            except:
+                return map(itertuple._make, zip(*arrays))
+            except Exception:
                 pass
 
         # fallback to regular tuples

pandas/tests/test_frame.py

@@ -5554,6 +5554,8 @@ def test_itertuples(self):
         dfaa = df[['a', 'a']]
         self.assertEqual(list(dfaa.itertuples()), [(0, 1, 1), (1, 2, 2), (2, 3, 3)])
 
+        self.assertEqual(repr(list(df.itertuples(name=None))), '[(0, 1, 4), (1, 2, 5), (2, 3, 6)]')
+
         tup = next(df.itertuples(name='TestName'))
 
         # no support for field renaming in Python 2.6, regular tuples are returned