Backport PR #45283: BUG: do not replace all nulls with "NaN"-string i…

…n Series index (#45473)

Co-authored-by: realead <egor.dranischnikow@googlemail.com>

doc/source/whatsnew/v1.4.0.rst

@@ -1050,6 +1050,7 @@ Other
 - Bug in :meth:`Series.replace` raising ``ValueError`` when using ``regex=True`` with a :class:`Series` containing ``np.nan`` values (:issue:`43344`)
 - Bug in :meth:`DataFrame.to_records` where an incorrect ``n`` was used when missing names were replaced by ``level_n`` (:issue:`44818`)
 - Bug in :meth:`DataFrame.eval` where ``resolvers`` argument was overriding the default resolvers (:issue:`34966`)
+- :meth:`Series.__repr__` and :meth:`DataFrame.__repr__` no longer replace all null-values in indexes with "NaN" but use their real string-representations. "NaN" is used only for ``float("nan")`` (:issue:`45263`)
 
 .. ---------------------------------------------------------------------------
 

pandas/_libs/missing.pyi

@@ -14,3 +14,4 @@ def checknull(val: object, inf_as_na: bool = ...) -> bool: ...
 def isnaobj(arr: np.ndarray, inf_as_na: bool = ...) -> npt.NDArray[np.bool_]: ...
 def isnaobj2d(arr: np.ndarray, inf_as_na: bool = ...) -> npt.NDArray[np.bool_]: ...
 def is_numeric_na(values: np.ndarray) -> npt.NDArray[np.bool_]: ...
+def is_float_nan(values: np.ndarray) -> npt.NDArray[np.bool_]: ...

pandas/_libs/missing.pyx

@@ -248,6 +248,31 @@ cdef bint checknull_with_nat_and_na(object obj):
     return checknull_with_nat(obj) or obj is C_NA
 
 
+@cython.wraparound(False)
+@cython.boundscheck(False)
+def is_float_nan(values: ndarray) -> ndarray:
+    """
+    True for elements which correspond to a float nan
+
+    Returns
+    -------
+    ndarray[bool]
+    """
+    cdef:
+        ndarray[uint8_t] result
+        Py_ssize_t i, N
+        object val
+
+    N = len(values)
+    result = np.zeros(N, dtype=np.uint8)
+
+    for i in range(N):
+        val = values[i]
+        if util.is_nan(val):
+            result[i] = True
+    return result.view(bool)
+
+
 @cython.wraparound(False)
 @cython.boundscheck(False)
 def is_numeric_na(values: ndarray) -> ndarray:

pandas/core/indexes/base.py

@@ -33,6 +33,7 @@
     is_datetime_array,
     no_default,
 )
+from pandas._libs.missing import is_float_nan
 from pandas._libs.tslibs import (
     IncompatibleFrequency,
     OutOfBoundsDatetime,
@@ -1391,7 +1392,7 @@ def _format_with_header(self, header: list[str_t], na_rep: str_t) -> list[str_t]
             result = [pprint_thing(x, escape_chars=("\t", "\r", "\n")) for x in values]
 
             # could have nans
-            mask = isna(values)
+            mask = is_float_nan(values)
             if mask.any():
                 result_arr = np.array(result)
                 result_arr[mask] = na_rep

pandas/tests/frame/test_repr_info.py

@@ -68,6 +68,27 @@ def test_repr_with_mi_nat(self, float_string_frame):
         expected = "              X\nNaT        a  1\n2013-01-01 b  2"
         assert result == expected
 
+    def test_repr_with_different_nulls(self):
+        # GH45263
+        df = DataFrame([1, 2, 3, 4], [True, None, np.nan, NaT])
+        result = repr(df)
+        expected = """      0
+True  1
+None  2
+NaN   3
+NaT   4"""
+        assert result == expected
+
+    def test_repr_with_different_nulls_cols(self):
+        # GH45263
+        d = {np.nan: [1, 2], None: [3, 4], NaT: [6, 7], True: [8, 9]}
+        df = DataFrame(data=d)
+        result = repr(df)
+        expected = """   NaN  None  NaT  True
+0    1     3    6     8
+1    2     4    7     9"""
+        assert result == expected
+
     def test_multiindex_na_repr(self):
         # only an issue with long columns
         df3 = DataFrame(

pandas/tests/indexes/test_base.py

@@ -658,7 +658,8 @@ def test_format_missing(self, vals, nulls_fixture):
         index = Index(vals)
 
         formatted = index.format()
-        expected = [str(index[0]), str(index[1]), str(index[2]), "NaN"]
+        null_repr = "NaN" if isinstance(nulls_fixture, float) else str(nulls_fixture)
+        expected = [str(index[0]), str(index[1]), str(index[2]), null_repr]
 
         assert formatted == expected
         assert index[3] is nulls_fixture

pandas/tests/series/test_repr.py

@@ -258,6 +258,13 @@ def test_float_repr(self):
         expected = "0    1.0\ndtype: object"
         assert repr(ser) == expected
 
+    def test_different_null_objects(self):
+        # GH#45263
+        ser = Series([1, 2, 3, 4], [True, None, np.nan, pd.NaT])
+        result = repr(ser)
+        expected = "True    1\nNone    2\nNaN     3\nNaT     4\ndtype: int64"
+        assert result == expected
+
 
 class TestCategoricalRepr:
     def test_categorical_repr_unicode(self):