DEPR: silently ignoring unrecognized timezones (#51477)

* DEPR: silently ignoring unrecognized timezones

* catch to_datetime cases

Author: jbrockmendel

doc/source/whatsnew/v2.0.0.rst

@@ -838,6 +838,7 @@ Other API changes
 Deprecations
 ~~~~~~~~~~~~
 - Deprecated parsing datetime strings with system-local timezone to ``tzlocal``, pass a ``tz`` keyword or explicitly call ``tz_localize`` instead (:issue:`50791`)
+- Deprecated silently dropping unrecognized timezones when parsing strings to datetimes (:issue:`18702`)
 - Deprecated argument ``infer_datetime_format`` in :func:`to_datetime` and :func:`read_csv`, as a strict version of it is now the default (:issue:`48621`)
 - Deprecated behavior of :func:`to_datetime` with ``unit`` when parsing strings, in a future version these will be parsed as datetimes (matching unit-less behavior) instead of cast to floats. To retain the old behavior, cast strings to numeric types before calling :func:`to_datetime` (:issue:`50735`)
 - Deprecated :func:`pandas.io.sql.execute` (:issue:`50185`)

pandas/_libs/tslibs/parsing.pyx

@@ -722,6 +722,18 @@ cdef datetime dateutil_parse(
                     f'Parsed string "{timestr}" gives an invalid tzoffset, '
                     "which must be between -timedelta(hours=24) and timedelta(hours=24)"
                 )
+        elif res.tzname is not None:
+            # e.g. "1994 Jan 15 05:16 FOO" where FOO is not recognized
+            # GH#18702
+            warnings.warn(
+                f'Parsed string "{timestr}" included an un-recognized timezone '
+                f'"{res.tzname}". Dropping unrecognized timezones is deprecated; '
+                "in a future version this will raise. Instead pass the string "
+                "without the timezone, then use .tz_localize to convert to a "
+                "recognized timezone.",
+                FutureWarning,
+                stacklevel=find_stack_level()
+            )
 
     out_bestunit[0] = attrname_to_npy_unit[reso]
     return ret
@@ -865,6 +877,8 @@ def guess_datetime_format(dt_str: str, bint dayfirst=False) -> str | None:
         datetime format string (for `strftime` or `strptime`),
         or None if it can't be guessed.
     """
+    cdef:
+        NPY_DATETIMEUNIT out_bestunit
     day_attribute_and_format = (("day",), "%d", 2)
 
     # attr name, format, padding (if any)
@@ -895,8 +909,17 @@ def guess_datetime_format(dt_str: str, bint dayfirst=False) -> str | None:
         datetime_attrs_to_format.remove(day_attribute_and_format)
         datetime_attrs_to_format.insert(0, day_attribute_and_format)
 
+    # same default used by dateutil
+    default = datetime.now().replace(hour=0, minute=0, second=0, microsecond=0)
     try:
-        parsed_datetime = du_parse(dt_str, dayfirst=dayfirst)
+        parsed_datetime = dateutil_parse(
+            dt_str,
+            default=default,
+            dayfirst=dayfirst,
+            yearfirst=False,
+            ignoretz=False,
+            out_bestunit=&out_bestunit,
+        )
     except (ValueError, OverflowError, InvalidOperation):
         # In case the datetime can't be parsed, its format cannot be guessed
         return None

pandas/tests/tools/test_to_datetime.py

@@ -3595,3 +3595,19 @@ def test_to_datetime_mixed_not_necessarily_iso8601_coerce(errors, expected):
     # https://github.com/pandas-dev/pandas/issues/50411
     result = to_datetime(["2020-01-01", "01-01-2000"], format="ISO8601", errors=errors)
     tm.assert_index_equal(result, expected)
+
+
+def test_ignoring_unknown_tz_deprecated():
+    # GH#18702, GH#51476
+    dtstr = "2014 Jan 9 05:15 FAKE"
+    msg = 'un-recognized timezone "FAKE". Dropping unrecognized timezones is deprecated'
+    with tm.assert_produces_warning(FutureWarning, match=msg):
+        res = Timestamp(dtstr)
+    assert res == Timestamp(dtstr[:-5])
+
+    with tm.assert_produces_warning(FutureWarning):
+        res = to_datetime(dtstr)
+    assert res == to_datetime(dtstr[:-5])
+    with tm.assert_produces_warning(FutureWarning):
+        res = to_datetime([dtstr])
+    tm.assert_index_equal(res, to_datetime([dtstr[:-5]]))