BUG: to_datetime raising on invalid offsets with errors=coerce and infer_datetime_format (#48676)

* BUG: to_datetime raising on invalid offsets with errors=coerce and infer_datetime_format

* fix invalid type test

* test errors=ignore as well

Co-authored-by: MarcoGorelli <>

Author: MarcoGorelli

doc/source/whatsnew/v1.6.0.rst

@@ -166,8 +166,10 @@ Categorical
 Datetimelike
 ^^^^^^^^^^^^
 - Bug in :func:`pandas.infer_freq`, raising ``TypeError`` when inferred on :class:`RangeIndex` (:issue:`47084`)
+- Bug in :func:`to_datetime` was raising on invalid offsets with ``errors='coerce'`` and ``infer_datetime_format=True`` (:issue:`48633`)
 - Bug in :class:`DatetimeIndex` constructor failing to raise when ``tz=None`` is explicitly specified in conjunction with timezone-aware ``dtype`` or data (:issue:`48659`)
 - Bug in subtracting a ``datetime`` scalar from :class:`DatetimeIndex` failing to retain the original ``freq`` attribute (:issue:`48818`)
+-
 
 Timedelta
 ^^^^^^^^^

pandas/_libs/tslibs/parsing.pyx

@@ -943,7 +943,7 @@ def format_is_iso(f: str) -> bint:
     return False
 
 
-def guess_datetime_format(dt_str, bint dayfirst=False):
+def guess_datetime_format(dt_str: str, bint dayfirst=False) -> str | None:
     """
     Guess the datetime format of a given datetime string.
 
@@ -1026,7 +1026,12 @@ def guess_datetime_format(dt_str, bint dayfirst=False):
             # This separation will prevent subsequent processing
             # from correctly parsing the time zone format.
             # So in addition to the format nomalization, we rejoin them here.
-            tokens[offset_index] = parsed_datetime.strftime("%z")
+            try:
+                tokens[offset_index] = parsed_datetime.strftime("%z")
+            except ValueError:
+                # Invalid offset might not have raised in du_parse
+                # https://github.com/dateutil/dateutil/issues/188
+                return None
             tokens = tokens[:offset_index + 1 or None]
 
     format_guess = [None] * len(tokens)

pandas/tests/tools/test_to_datetime.py

@@ -1142,13 +1142,37 @@ def test_to_datetime_coerce(self):
         )
         tm.assert_index_equal(result, expected)
 
-    def test_to_datetime_coerce_malformed(self):
+    @pytest.mark.parametrize("infer_datetime_format", [True, False])
+    @pytest.mark.parametrize(
+        "errors, expected",
+        [
+            ("coerce", Index([NaT, NaT])),
+            ("ignore", Index(["200622-12-31", "111111-24-11"])),
+        ],
+    )
+    def test_to_datetime_malformed_no_raise(
+        self, errors, expected, infer_datetime_format
+    ):
         # GH 28299
+        # GH 48633
         ts_strings = ["200622-12-31", "111111-24-11"]
-        result = to_datetime(ts_strings, errors="coerce")
-        expected = Index([NaT, NaT])
+        result = to_datetime(
+            ts_strings, errors=errors, infer_datetime_format=infer_datetime_format
+        )
         tm.assert_index_equal(result, expected)
 
+    @pytest.mark.parametrize("infer_datetime_format", [True, False])
+    def test_to_datetime_malformed_raise(self, infer_datetime_format):
+        # GH 48633
+        ts_strings = ["200622-12-31", "111111-24-11"]
+        with pytest.raises(
+            ValueError,
+            match=r"^hour must be in 0\.\.23: 111111-24-11 present at position 1$",
+        ):
+            to_datetime(
+                ts_strings, errors="raise", infer_datetime_format=infer_datetime_format
+            )
+
     def test_iso_8601_strings_with_same_offset(self):
         # GH 17697, 11736
         ts_str = "2015-11-18 15:30:00+05:30"

pandas/tests/tslibs/test_parsing.py

@@ -212,8 +212,6 @@ def test_guess_datetime_format_with_locale_specific_formats(string, fmt):
         "1/1/1/1",
         "this_is_not_a_datetime",
         "51a",
-        9,
-        datetime(2011, 1, 1),
     ],
 )
 def test_guess_datetime_format_invalid_inputs(invalid_dt):
@@ -222,6 +220,17 @@ def test_guess_datetime_format_invalid_inputs(invalid_dt):
     assert parsing.guess_datetime_format(invalid_dt) is None
 
 
+@pytest.mark.parametrize("invalid_type_dt", [9, datetime(2011, 1, 1)])
+def test_guess_datetime_format_wrong_type_inputs(invalid_type_dt):
+    # A datetime string must include a year, month and a day for it to be
+    # guessable, in addition to being a string that looks like a datetime.
+    with pytest.raises(
+        TypeError,
+        match=r"^Argument 'dt_str' has incorrect type \(expected str, got .*\)$",
+    ):
+        parsing.guess_datetime_format(invalid_type_dt)
+
+
 @pytest.mark.parametrize(
     "string,fmt",
     [