check for nat_strings when finding first null

Author: MarcoGorelli

pandas/_libs/tslib.pyx

@@ -421,20 +421,23 @@ def array_with_unit_to_datetime(
 
     return oresult, tz
 
-def first_non_null(values: ndarray):
-    """Find first non-null value, return None if there isn't one."""
+@cython.wraparound(False)
+@cython.boundscheck(False)
+def first_non_null(values: ndarray) -> int:
+    """Find position of first non-null value, return -1 if there isn't one."""
     cdef:
         Py_ssize_t n = len(values)
         Py_ssize_t i
+        int result
     for i in range(n):
         val = values[i]
         if checknull_with_nat_and_na(val):
             continue
         if isinstance(val, str) and (len(val) == 0 or val in nat_strings):
             continue
-        return val
+        return i
     else:
-        return None
+        return -1
 
 @cython.wraparound(False)
 @cython.boundscheck(False)

pandas/core/tools/datetimes.py

@@ -129,16 +129,19 @@ class FulldatetimeDict(YearMonthDayDict, total=False):
 
 def _guess_datetime_format_for_array(arr, dayfirst: bool | None = False) -> str | None:
     # Try to guess the format based on the first non-NaN element, return None if can't
-    if type(first_non_nan_element := tslib.first_non_null(arr)) is str:
-        # GH#32264 np.str_ object
-        guessed_format = guess_datetime_format(first_non_nan_element, dayfirst=dayfirst)
-        if guessed_format is not None:
-            return guessed_format
-        warnings.warn(
-            "Could not infer format - "
-            "to ensure consistent parsing, specify a format.",
-            stacklevel=find_stack_level(),
-        )
+    if (first_non_null := tslib.first_non_null(arr)) != -1:
+        if type(first_non_nan_element := arr[first_non_null]) is str:
+            # GH#32264 np.str_ object
+            guessed_format = guess_datetime_format(
+                first_non_nan_element, dayfirst=dayfirst
+            )
+            if guessed_format is not None:
+                return guessed_format
+            warnings.warn(
+                "Could not infer format - "
+                "to ensure consistent parsing, specify a format.",
+                stacklevel=find_stack_level(),
+            )
     return None
 
 

pandas/tests/tools/test_to_datetime.py

@@ -2100,9 +2100,9 @@ def test_to_datetime_dta_tz(self, klass):
 
 
 class TestGuessDatetimeFormat:
-    # @td.skip_if_not_us_locale
+    @td.skip_if_not_us_locale
     @pytest.mark.parametrize(
-        "test_array",
+        "test_list",
         [
             [
                 "2011-12-30 00:00:00.000000",
@@ -2115,8 +2115,9 @@ class TestGuessDatetimeFormat:
             ["2011-12-30 00:00:00.000000", "random_string"],
         ],
     )
-    def test_guess_datetime_format_for_array(self, test_array):
+    def test_guess_datetime_format_for_array(self, test_list):
         expected_format = "%Y-%m-%d %H:%M:%S.%f"
+        test_array = np.array(test_list, dtype=object)
         assert tools._guess_datetime_format_for_array(test_array) == expected_format
 
     @td.skip_if_not_us_locale