Fixed BUG: Using NamedTuples with .iloc fails #48188

pandas/core/indexing.py

@@ -1162,7 +1162,8 @@ def _check_deprecated_callable_usage(self, key: Any, maybe_callable: T) -> T:
     @final
     def __getitem__(self, key):
         check_dict_or_set_indexers(key)
-        if type(key) is tuple:
+        ##if type(key) is tuple:
+        if isinstance(key, tuple):
             key = tuple(list(x) if is_iterator(x) else x for x in key)
             key = tuple(com.apply_if_callable(x, self.obj) for x in key)
             if self._is_scalar_access(key):

pandas/tests/indexing/test_iloc.py

@@ -6,6 +6,10 @@
 import numpy as np
 import pytest
 
+
+import pandas as pd
+from collections import namedtuple
+
 from pandas.errors import IndexingError
 
 from pandas import (
@@ -1463,3 +1467,24 @@ def test_iloc_nullable_int64_size_1_nan(self):
             result.loc[:, "b"] = result.loc[:, "b"].astype("Int64")
         expected = DataFrame({"a": ["test"], "b": array([NA], dtype="Int64")})
         tm.assert_frame_equal(result, expected)
+
+
+
+    def test_iloc_with_namedtuple():
+        # setup a dataFrame
+        df = pd.DataFrame(np.arange(100).reshape(10, 10))
+
+        # define a namedtuple 
+        Indexer = namedtuple("Indexer", ["row", "col"])
+
+        # Test .iloc with a regular tuple
+        result_with_tuple = df.iloc[(1, 2)]
+        assert result_with_tuple == df.iloc[1, 2]
+
+        # Test .iloc with a namedtuple
+        indexer = Indexer(row=1, col=2)
+        result_with_namedtuple = df.iloc[indexer]
+        assert result_with_namedtuple == df.iloc[1, 2]
+
+        # ensure both methods yield the same result testing the initial issue
+        assert result_with_tuple == result_with_namedtuple