PERF: optimize is_scalar, is_iterator

pandas/_libs/lib.pyx

@@ -11,6 +11,9 @@ from cython import Py_ssize_t
 from cpython.object cimport PyObject_RichCompareBool, Py_EQ
 from cpython.ref cimport Py_INCREF
 from cpython.tuple cimport PyTuple_SET_ITEM, PyTuple_New
+from cpython.iterator cimport PyIter_Check
+from cpython.sequence cimport PySequence_Check
+from cpython.number cimport PyNumber_Check
 
 from cpython.datetime cimport (PyDateTime_Check, PyDate_Check,
                                PyTime_Check, PyDelta_Check,
@@ -156,22 +159,63 @@ def is_scalar(val: object) -> bool:
     True
     """
 
-    return (cnp.PyArray_IsAnyScalar(val)
+    # Start with C-optimized checks
+    if (cnp.PyArray_IsAnyScalar(val)
             # PyArray_IsAnyScalar is always False for bytearrays on Py3
             or PyDate_Check(val)
             or PyDelta_Check(val)
             or PyTime_Check(val)
             # We differ from numpy, which claims that None is not scalar;
             # see np.isscalar
             or val is C_NA
-            or val is None
-            or isinstance(val, (Fraction, Number))
+            or val is None):
+        return True
+
+    # Next use C-optimized checks to exclude common non-scalars before falling
+    #  back to non-optimized checks.
+    if PySequence_Check(val):
+        # e.g. list, tuple
+        # includes np.ndarray, Series which PyNumber_Check can return True for
+        return False
+
+    # Note: PyNumber_Check check includes Decimal, Fraction, numbers.Number
+    return (PyNumber_Check(val)
             or util.is_period_object(val)
-            or is_decimal(val)
             or is_interval(val)
             or util.is_offset_object(val))
 
 
+def is_iterator(obj: object) -> bool:
+    """
+    Check if the object is an iterator.
+
+    This is intended for generators, not list-like objects.
+
+    Parameters
+    ----------
+    obj : The object to check
+
+    Returns
+    -------
+    is_iter : bool
+        Whether `obj` is an iterator.
+
+    Examples
+    --------
+    >>> is_iterator((x for x in []))
+    True
+    >>> is_iterator([1, 2, 3])
+    False
+    >>> is_iterator(datetime(2017, 1, 1))
+    False
+    >>> is_iterator("foo")
+    False
+    >>> is_iterator(1)
+    False
+    """
+    return PyIter_Check(obj)
+
+
 def item_from_zerodim(val: object) -> object:
     """
     If the value is a zerodim array, return the item it contains.

pandas/core/dtypes/inference.py

@@ -25,6 +25,8 @@
 
 is_list_like = lib.is_list_like
 
+is_iterator = lib.is_iterator
+
 
 def is_number(obj) -> bool:
     """
@@ -93,40 +95,6 @@ def _iterable_not_string(obj) -> bool:
     return isinstance(obj, abc.Iterable) and not isinstance(obj, str)
 
 
-def is_iterator(obj) -> bool:
-    """
-    Check if the object is an iterator.
-
-    For example, lists are considered iterators
-    but not strings or datetime objects.
-
-    Parameters
-    ----------
-    obj : The object to check
-
-    Returns
-    -------
-    is_iter : bool
-        Whether `obj` is an iterator.
-
-    Examples
-    --------
-    >>> is_iterator([1, 2, 3])
-    True
-    >>> is_iterator(datetime(2017, 1, 1))
-    False
-    >>> is_iterator("foo")
-    False
-    >>> is_iterator(1)
-    False
-    """
-
-    if not hasattr(obj, "__iter__"):
-        return False
-
-    return hasattr(obj, "__next__")
-
-
 def is_file_like(obj) -> bool:
     """
     Check if the object is a file-like object.

pandas/tests/dtypes/test_inference.py

@@ -1346,9 +1346,11 @@ def test_is_scalar_builtin_scalars(self):
         assert is_scalar(None)
         assert is_scalar(True)
         assert is_scalar(False)
-        assert is_scalar(Number())
         assert is_scalar(Fraction())
         assert is_scalar(0.0)
+        assert is_scalar(1)
+        assert is_scalar(complex(2))
+        assert is_scalar(float("NaN"))
         assert is_scalar(np.nan)
         assert is_scalar("foobar")
         assert is_scalar(b"foobar")
@@ -1357,6 +1359,7 @@ def test_is_scalar_builtin_scalars(self):
         assert is_scalar(time(12, 0))
         assert is_scalar(timedelta(hours=1))
         assert is_scalar(pd.NaT)
+        assert is_scalar(pd.NA)
 
     def test_is_scalar_builtin_nonscalars(self):
         assert not is_scalar({})
@@ -1371,6 +1374,7 @@ def test_is_scalar_numpy_array_scalars(self):
         assert is_scalar(np.int64(1))
         assert is_scalar(np.float64(1.0))
         assert is_scalar(np.int32(1))
+        assert is_scalar(np.complex64(2))
         assert is_scalar(np.object_("foobar"))
         assert is_scalar(np.str_("foobar"))
         assert is_scalar(np.unicode_("foobar"))
@@ -1410,6 +1414,21 @@ def test_is_scalar_pandas_containers(self):
         assert not is_scalar(Index([]))
         assert not is_scalar(Index([1]))
 
+    def test_is_scalar_number(self):
+        # Number() is not recognied by PyNumber_Check, so by extension
+        #  is not recognized by is_scalar, but instances of non-abstract
+        #  subclasses are.
+
+        class Numeric(Number):
+            def __init__(self, value):
+                self.value = value
+
+            def __int__(self):
+                return self.value
+
+        num = Numeric(1)
+        assert is_scalar(num)
+
 
 def test_datetimeindex_from_empty_datetime64_array():
     for unit in ["ms", "us", "ns"]: