fix: dbdate and dbtime support set item with null values (#85)

feat: dbdate and dbtime support numpy.datetime64 values in array constructor

Author: tswast

db_dtypes/__init__.py

@@ -106,6 +106,9 @@ def _datetime(
             r"(?:\.(?P<fraction>\d*))?)?)?\s*$"
         ).match,
     ) -> Optional[numpy.datetime64]:
+        if isinstance(scalar, numpy.datetime64):
+            return scalar
+
         # Convert pyarrow values to datetime.time.
         if isinstance(scalar, (pyarrow.Time32Scalar, pyarrow.Time64Scalar)):
             scalar = (
@@ -116,7 +119,7 @@ def _datetime(
             )
 
         if pandas.isna(scalar):
-            return None
+            return numpy.datetime64("NaT")
         if isinstance(scalar, datetime.time):
             return pandas.Timestamp(
                 year=1970,
@@ -238,12 +241,15 @@ def _datetime(
         scalar,
         match_fn=re.compile(r"\s*(?P<year>\d+)-(?P<month>\d+)-(?P<day>\d+)\s*$").match,
     ) -> Optional[numpy.datetime64]:
+        if isinstance(scalar, numpy.datetime64):
+            return scalar
+
         # Convert pyarrow values to datetime.date.
         if isinstance(scalar, (pyarrow.Date32Scalar, pyarrow.Date64Scalar)):
             scalar = scalar.as_py()
 
         if pandas.isna(scalar):
-            return None
+            return numpy.datetime64("NaT")
         elif isinstance(scalar, datetime.date):
             return pandas.Timestamp(
                 year=scalar.year, month=scalar.month, day=scalar.day

db_dtypes/core.py

@@ -100,14 +100,6 @@ def _cmp_method(self, other, op):
             return NotImplemented
         return op(self._ndarray, other._ndarray)
 
-    def __setitem__(self, key, value):
-        if is_list_like(value):
-            _datetime = self._datetime
-            value = [_datetime(v) for v in value]
-        elif not pandas.isna(value):
-            value = self._datetime(value)
-        return super().__setitem__(key, value)
-
     def _from_factorized(self, unique, original):
         return self.__class__(unique)
 
@@ -121,6 +113,16 @@ def _validate_scalar(self, value):
         """
         return self._datetime(value)
 
+    def _validate_setitem_value(self, value):
+        """
+        Convert a value for use in setting a value in the backing numpy array.
+        """
+        if is_list_like(value):
+            _datetime = self._datetime
+            return [_datetime(v) for v in value]
+
+        return self._datetime(value)
+
     def any(
         self,
         *,

db_dtypes/pandas_backports.py

@@ -126,7 +126,7 @@ def __getitem__(self, index):
         return self.__class__(value, self._dtype)
 
     def __setitem__(self, index, value):
-        self._ndarray[index] = value
+        self._ndarray[index] = self._validate_setitem_value(value)
 
     def __len__(self):
         return len(self._ndarray)

tests/unit/test_date.py

@@ -24,6 +24,33 @@
 from db_dtypes import pandas_backports
 
 
+VALUE_PARSING_TEST_CASES = [
+    # Min/Max values for pandas.Timestamp.
+    ("1677-09-22", datetime.date(1677, 9, 22)),
+    ("2262-04-11", datetime.date(2262, 4, 11)),
+    # Typical "zero" values.
+    ("1900-01-01", datetime.date(1900, 1, 1)),
+    ("1970-01-01", datetime.date(1970, 1, 1)),
+    # Assorted values.
+    ("1993-10-31", datetime.date(1993, 10, 31)),
+    (datetime.date(1993, 10, 31), datetime.date(1993, 10, 31)),
+    ("2012-02-29", datetime.date(2012, 2, 29)),
+    (numpy.datetime64("2012-02-29"), datetime.date(2012, 2, 29)),
+    ("2021-12-17", datetime.date(2021, 12, 17)),
+    (pandas.Timestamp("2021-12-17"), datetime.date(2021, 12, 17)),
+    ("2038-01-19", datetime.date(2038, 1, 19)),
+]
+
+NULL_VALUE_TEST_CASES = [
+    None,
+    pandas.NaT,
+    float("nan"),
+]
+
+if hasattr(pandas, "NA"):
+    NULL_VALUE_TEST_CASES.append(pandas.NA)
+
+
 def test_box_func():
     input_array = db_dtypes.DateArray([])
     input_datetime = datetime.datetime(2022, 3, 16)
@@ -58,26 +85,49 @@ def test__cmp_method_with_scalar():
     assert got[0]
 
 
-@pytest.mark.parametrize(
-    "value, expected",
-    [
-        # Min/Max values for pandas.Timestamp.
-        ("1677-09-22", datetime.date(1677, 9, 22)),
-        ("2262-04-11", datetime.date(2262, 4, 11)),
-        # Typical "zero" values.
-        ("1900-01-01", datetime.date(1900, 1, 1)),
-        ("1970-01-01", datetime.date(1970, 1, 1)),
-        # Assorted values.
-        ("1993-10-31", datetime.date(1993, 10, 31)),
-        ("2012-02-29", datetime.date(2012, 2, 29)),
-        ("2021-12-17", datetime.date(2021, 12, 17)),
-        ("2038-01-19", datetime.date(2038, 1, 19)),
-    ],
-)
+@pytest.mark.parametrize("value, expected", VALUE_PARSING_TEST_CASES)
 def test_date_parsing(value, expected):
     assert pandas.Series([value], dtype="dbdate")[0] == expected
 
 
+@pytest.mark.parametrize("value", NULL_VALUE_TEST_CASES)
+def test_date_parsing_null(value):
+    assert pandas.Series([value], dtype="dbdate")[0] is pandas.NaT
+
+
+@pytest.mark.parametrize("value, expected", VALUE_PARSING_TEST_CASES)
+def test_date_set_item(value, expected):
+    series = pandas.Series([None], dtype="dbdate")
+    series[0] = value
+    assert series[0] == expected
+
+
+@pytest.mark.parametrize("value", NULL_VALUE_TEST_CASES)
+def test_date_set_item_null(value):
+    series = pandas.Series(["1970-01-01"], dtype="dbdate")
+    series[0] = value
+    assert series[0] is pandas.NaT
+
+
+def test_date_set_slice():
+    series = pandas.Series([None, None, None], dtype="dbdate")
+    series[:] = [
+        datetime.date(2022, 3, 21),
+        "2011-12-13",
+        numpy.datetime64("1998-09-04"),
+    ]
+    assert series[0] == datetime.date(2022, 3, 21)
+    assert series[1] == datetime.date(2011, 12, 13)
+    assert series[2] == datetime.date(1998, 9, 4)
+
+
+def test_date_set_slice_null():
+    series = pandas.Series(["1970-01-01"] * len(NULL_VALUE_TEST_CASES), dtype="dbdate")
+    series[:] = NULL_VALUE_TEST_CASES
+    for row_index in range(len(NULL_VALUE_TEST_CASES)):
+        assert series[row_index] is pandas.NaT
+
+
 @pytest.mark.parametrize(
     "value, error",
     [

tests/unit/test_time.py

@@ -73,8 +73,14 @@ def test_box_func():
         # Fractional seconds can cause rounding problems if cast to float. See:
         # https://github.com/googleapis/python-db-dtypes-pandas/issues/18
         ("0:0:59.876543", datetime.time(0, 0, 59, 876543)),
+        (
+            numpy.datetime64("1970-01-01 00:00:59.876543"),
+            datetime.time(0, 0, 59, 876543),
+        ),
         ("01:01:01.010101", datetime.time(1, 1, 1, 10101)),
+        (pandas.Timestamp("1970-01-01 01:01:01.010101"), datetime.time(1, 1, 1, 10101)),
         ("09:09:09.090909", datetime.time(9, 9, 9, 90909)),
+        (datetime.time(9, 9, 9, 90909), datetime.time(9, 9, 9, 90909)),
         ("11:11:11.111111", datetime.time(11, 11, 11, 111111)),
         ("19:16:23.987654", datetime.time(19, 16, 23, 987654)),
         # Microsecond precision