changes based on comments on PR

Author: sshu2017

doc/source/whatsnew/v3.0.0.rst

@@ -618,6 +618,7 @@ I/O
 - Bug in :meth:`read_csv` causing segmentation fault when ``encoding_errors`` is not a string. (:issue:`59059`)
 - Bug in :meth:`read_csv` raising ``TypeError`` when ``index_col`` is specified and ``na_values`` is a dict containing the key ``None``. (:issue:`57547`)
 - Bug in :meth:`read_csv` raising ``TypeError`` when ``nrows`` and ``iterator`` are specified without specifying a ``chunksize``. (:issue:`59079`)
+- Bug in :meth:`read_csv` where the order of the ``na_values`` makes an inconsistency when ``na_values`` is a list non-string values. (:issue:`59303`)
 - Bug in :meth:`read_excel` raising ``ValueError`` when passing array of boolean values when ``dtype="boolean"``. (:issue:`58159`)
 - Bug in :meth:`read_json` not validating the ``typ`` argument to not be exactly ``"frame"`` or ``"series"`` (:issue:`59124`)
 - Bug in :meth:`read_stata` raising ``KeyError`` when input file is stored in big-endian format and contains strL data. (:issue:`58638`)
@@ -693,7 +694,6 @@ Other
 - Bug in :meth:`Series.dt` methods in :class:`ArrowDtype` that were returning incorrect values. (:issue:`57355`)
 - Bug in :meth:`Series.rank` that doesn't preserve missing values for nullable integers when ``na_option='keep'``. (:issue:`56976`)
 - Bug in :meth:`Series.replace` and :meth:`DataFrame.replace` inconsistently replacing matching instances when ``regex=True`` and missing values are present. (:issue:`56599`)
-- Bug in :meth:`_clean_na_values` in :class:`TextFileReader` that was not properly handling ``na_values`` when it is a list of strings. (:issue:`59303`)
 - Bug in Dataframe Interchange Protocol implementation was returning incorrect results for data buffers' associated dtype, for string and datetime columns (:issue:`54781`)
 - Bug in ``Series.list`` methods not preserving the original :class:`Index`. (:issue:`58425`)
 

pandas/tests/io/parser/test_na_values.py

@@ -815,20 +815,21 @@ def test_bool_and_nan_to_float(all_parsers):
 
 
 @xfail_pyarrow
-def test_na_values_dict_without_dtype(all_parsers):
-    # GH#59303
+@pytest.mark.parametrize(
+    "na_values, expected_result, test_id",
+    [
+        ({"A": [-99.0, -99]}, DataFrame({"A": [np.nan, np.nan, np.nan, np.nan]}), "float_first"),
+        ({"A": [-99, -99.0]}, DataFrame({"A": [np.nan, np.nan, np.nan, np.nan]}), "int_first"),
+    ],
+    ids=["float_first", "int_first"]
+)
+def test_na_values_dict_without_dtype(all_parsers, na_values, expected_result, test_id):
     parser = all_parsers
     data = """A
 -99
 -99
 -99.0
 -99.0"""
-    # this would FAIL BEFORE this fix
-    result_1 = parser.read_csv(StringIO(data), na_values={"A": [-99.0, -99]})
-    expected_1 = DataFrame.from_dict({"A": [np.nan, np.nan, np.nan, np.nan]})
-    tm.assert_frame_equal(result_1, expected_1)
-
-    # this would PASS even BEFORE this fix
-    result_2 = parser.read_csv(StringIO(data), na_values={"A": [-99, -99.0]})
-    expected_2 = DataFrame.from_dict({"A": [np.nan, np.nan, np.nan, np.nan]})
-    tm.assert_frame_equal(result_2, expected_2)
+
+    result = parser.read_csv(StringIO(data), na_values=na_values)
+    tm.assert_frame_equal(result, expected_result)