Resolving issue request pandas-dev#29738. Updated integer.py to analyze the types on iteratively going through the collection to cast to the appropriate type. In addition added a basic test case test_native_calls_types to assert that the change works

Author: Brosinc132

pandas/core/arrays/integer.py

@@ -450,7 +450,7 @@ def __iter__(self):
             if self._mask[i]:
                 yield self.dtype.na_value
             else:
-                yield self._data[i]
+                yield self._data[i].item()
 
     def take(self, indexer, allow_fill=False, fill_value=None):
         # we always fill with 1 internally

pandas/tests/arrays/test_integer.py

@@ -860,6 +860,16 @@ def test_stat_method(pandasmethname, kwargs):
     expected = pandasmeth(**kwargs)
     assert expected == result
 
+def test_native_calls_types():
+    s = pd.Series([1, 2], dtype='int64')
+    assert (type(list(s.iteritems())[0][1]) == type(1))
+    assert(type(s[0]) != type(list(s.iteritems())[0][1]))
+    s = pd.Series([1, 2], dtype='Int64')
+    assert (type(list(s.iteritems())[0][1]) == type(1))
+
+
+
+
 
 # TODO(jreback) - these need testing / are broken