TST: dt64 unit in tests (#56126)

Author: jbrockmendel

pandas/core/tools/datetimes.py

@@ -495,7 +495,8 @@ def _convert_listlike_datetimes(
     if tz_parsed is not None:
         # We can take a shortcut since the datetime64 numpy array
         # is in UTC
-        dtype = cast(DatetimeTZDtype, tz_to_dtype(tz_parsed))
+        out_unit = np.datetime_data(result.dtype)[0]
+        dtype = cast(DatetimeTZDtype, tz_to_dtype(tz_parsed, out_unit))
         dt64_values = result.view(f"M8[{dtype.unit}]")
         dta = DatetimeArray._simple_new(dt64_values, dtype=dtype)
         return DatetimeIndex._simple_new(dta, name=name)

pandas/tests/arithmetic/test_timedelta64.py

@@ -1039,7 +1039,7 @@ def test_td64arr_add_datetime64_nat(self, box_with_array):
         other = np.datetime64("NaT")
 
         tdi = timedelta_range("1 day", periods=3)
-        expected = DatetimeIndex(["NaT", "NaT", "NaT"])
+        expected = DatetimeIndex(["NaT", "NaT", "NaT"], dtype="M8[ns]")
 
         tdser = tm.box_expected(tdi, box_with_array)
         expected = tm.box_expected(expected, box_with_array)

pandas/tests/dtypes/test_dtypes.py

@@ -200,7 +200,8 @@ def test_is_boolean(self, categories, expected):
 
     def test_dtype_specific_categorical_dtype(self):
         expected = "datetime64[ns]"
-        result = str(Categorical(DatetimeIndex([])).categories.dtype)
+        dti = DatetimeIndex([], dtype=expected)
+        result = str(Categorical(dti).categories.dtype)
         assert result == expected
 
     def test_not_string(self):

pandas/tests/frame/methods/test_transpose.py

@@ -29,7 +29,8 @@ def test_transpose_td64_intervals(self):
 
     def test_transpose_empty_preserves_datetimeindex(self):
         # GH#41382
-        df = DataFrame(index=DatetimeIndex([]))
+        dti = DatetimeIndex([], dtype="M8[ns]")
+        df = DataFrame(index=dti)
 
         expected = DatetimeIndex([], dtype="datetime64[ns]", freq=None)
 

pandas/tests/indexes/datetimes/test_date_range.py

@@ -183,6 +183,7 @@ def test_date_range_edges(self, freq):
         )
         exp = DatetimeIndex(
             [ts + n * td for n in range(1, 5)],
+            dtype="M8[ns]",
             freq=freq,
         )
         tm.assert_index_equal(idx, exp)
@@ -193,7 +194,7 @@ def test_date_range_edges(self, freq):
             end=ts + td,
             freq=freq,
         )
-        exp = DatetimeIndex([], freq=freq)
+        exp = DatetimeIndex([], dtype="M8[ns]", freq=freq)
         tm.assert_index_equal(idx, exp)
 
         # start matches end
@@ -202,7 +203,7 @@ def test_date_range_edges(self, freq):
             end=ts + td,
             freq=freq,
         )
-        exp = DatetimeIndex([ts + td], freq=freq)
+        exp = DatetimeIndex([ts + td], dtype="M8[ns]", freq=freq)
         tm.assert_index_equal(idx, exp)
 
     def test_date_range_near_implementation_bound(self):

pandas/tests/indexes/datetimes/test_scalar_compat.py

@@ -304,7 +304,7 @@ def test_dti_fields(self, tz):
         exp = dti[[0, 90, 181, 273]]
         tm.assert_index_equal(res, exp)
         res = dti[dti.is_leap_year]
-        exp = DatetimeIndex([], freq="D", tz=dti.tz, name="name")
+        exp = DatetimeIndex([], freq="D", tz=dti.tz, name="name").as_unit("ns")
         tm.assert_index_equal(res, exp)
 
     def test_dti_is_year_quarter_start(self):

pandas/tests/indexes/datetimes/test_setops.py

@@ -73,7 +73,7 @@ def test_union(self, tz, sort):
         expected2_notsorted = DatetimeIndex(list(other2) + list(rng2[:3]))
 
         rng3 = date_range("1/1/2000", freq="D", periods=5, tz=tz)
-        other3 = DatetimeIndex([], tz=tz)
+        other3 = DatetimeIndex([], tz=tz).as_unit("ns")
         expected3 = date_range("1/1/2000", freq="D", periods=5, tz=tz)
         expected3_notsorted = rng3
 
@@ -235,7 +235,7 @@ def test_intersection(self, tz, sort):
         expected3 = date_range("6/1/2000", "6/20/2000", freq="D", name=None)
 
         rng4 = date_range("7/1/2000", "7/31/2000", freq="D", name="idx")
-        expected4 = DatetimeIndex([], freq="D", name="idx")
+        expected4 = DatetimeIndex([], freq="D", name="idx", dtype="M8[ns]")
 
         for rng, expected in [
             (rng2, expected2),
@@ -265,7 +265,7 @@ def test_intersection(self, tz, sort):
 
         # GH 7880
         rng4 = date_range("7/1/2000", "7/31/2000", freq="D", tz=tz, name="idx")
-        expected4 = DatetimeIndex([], tz=tz, name="idx")
+        expected4 = DatetimeIndex([], tz=tz, name="idx").as_unit("ns")
         assert expected4.freq is None
 
         for rng, expected in [
@@ -536,7 +536,7 @@ def test_intersection(self):
 
         # non-overlapping
         the_int = rng[:10].intersection(rng[10:])
-        expected = DatetimeIndex([])
+        expected = DatetimeIndex([]).as_unit("ns")
         tm.assert_index_equal(the_int, expected)
 
     def test_intersection_bug(self):

pandas/tests/indexes/interval/test_interval.py

@@ -388,7 +388,7 @@ def test_maybe_convert_i8_nat(self, breaks):
         # GH 20636
         index = IntervalIndex.from_breaks(breaks)
 
-        to_convert = breaks._constructor([pd.NaT] * 3)
+        to_convert = breaks._constructor([pd.NaT] * 3).as_unit("ns")
         expected = Index([np.nan] * 3, dtype=np.float64)
         result = index._maybe_convert_i8(to_convert)
         tm.assert_index_equal(result, expected)

pandas/tests/indexes/period/test_partial_slicing.py

@@ -117,7 +117,7 @@ def test_range_slice_outofbounds(self, make_range):
         idx = make_range(start="2013/10/01", freq="D", periods=10)
 
         df = DataFrame({"units": [100 + i for i in range(10)]}, index=idx)
-        empty = DataFrame(index=type(idx)([], freq="D"), columns=["units"])
+        empty = DataFrame(index=idx[:0], columns=["units"])
         empty["units"] = empty["units"].astype("int64")
 
         tm.assert_frame_equal(df["2013/09/01":"2013/09/30"], empty)

pandas/tests/io/formats/test_format.py

@@ -872,20 +872,22 @@ def test_to_string_truncate_multilevel(self):
         with option_context("display.max_rows", 7, "display.max_columns", 7):
             assert has_doubly_truncated_repr(df)
 
-    def test_truncate_with_different_dtypes(self):
+    @pytest.mark.parametrize("dtype", ["object", "datetime64[us]"])
+    def test_truncate_with_different_dtypes(self, dtype):
         # 11594, 12045
         # when truncated the dtypes of the splits can differ
 
         # 11594
-        s = Series(
+        ser = Series(
             [datetime(2012, 1, 1)] * 10
             + [datetime(1012, 1, 2)]
-            + [datetime(2012, 1, 3)] * 10
+            + [datetime(2012, 1, 3)] * 10,
+            dtype=dtype,
         )
 
         with option_context("display.max_rows", 8):
-            result = str(s)
-            assert "object" in result
+            result = str(ser)
+        assert dtype in result
 
     def test_truncate_with_different_dtypes2(self):
         # 12045

pandas/tests/io/parser/test_parse_dates.py

@@ -831,9 +831,7 @@ def test_parse_dates_string(all_parsers):
     parser = all_parsers
     result = parser.read_csv(StringIO(data), index_col="date", parse_dates=["date"])
     # freq doesn't round-trip
-    index = DatetimeIndex(
-        list(date_range("1/1/2009", periods=3)), name="date", freq=None
-    )
+    index = date_range("1/1/2009", periods=3, name="date")._with_freq(None)
 
     expected = DataFrame(
         {"A": ["a", "b", "c"], "B": [1, 3, 4], "C": [2, 4, 5]}, index=index
@@ -1736,17 +1734,12 @@ def test_parse_timezone(all_parsers):
               2018-01-04 09:05:00+09:00,23400"""
     result = parser.read_csv(StringIO(data), parse_dates=["dt"])
 
-    dti = DatetimeIndex(
-        list(
-            date_range(
-                start="2018-01-04 09:01:00",
-                end="2018-01-04 09:05:00",
-                freq="1min",
-                tz=timezone(timedelta(minutes=540)),
-            )
-        ),
-        freq=None,
-    )
+    dti = date_range(
+        start="2018-01-04 09:01:00",
+        end="2018-01-04 09:05:00",
+        freq="1min",
+        tz=timezone(timedelta(minutes=540)),
+    )._with_freq(None)
     expected_data = {"dt": dti, "val": [23350, 23400, 23400, 23400, 23400]}
 
     expected = DataFrame(expected_data)

pandas/tests/resample/test_period_index.py

@@ -504,8 +504,8 @@ def test_resample_weekly_all_na(self):
         expected = ts.asfreq("W-THU").ffill()
         tm.assert_series_equal(result, expected)
 
-    def test_resample_tz_localized(self):
-        dr = date_range(start="2012-4-13", end="2012-5-1")
+    def test_resample_tz_localized(self, unit):
+        dr = date_range(start="2012-4-13", end="2012-5-1", unit=unit)
         ts = Series(range(len(dr)), index=dr)
 
         ts_utc = ts.tz_localize("UTC")
@@ -514,9 +514,7 @@ def test_resample_tz_localized(self):
         result = ts_local.resample("W").mean()
 
         ts_local_naive = ts_local.copy()
-        ts_local_naive.index = [
-            x.replace(tzinfo=None) for x in ts_local_naive.index.to_pydatetime()
-        ]
+        ts_local_naive.index = ts_local_naive.index.tz_localize(None)
 
         exp = ts_local_naive.resample("W").mean().tz_localize("America/Los_Angeles")
         exp.index = pd.DatetimeIndex(exp.index, freq="W")

pandas/tests/resample/test_resampler_grouper.py

@@ -349,18 +349,19 @@ def weighted_quantile(series, weights, q):
     tm.assert_series_equal(result, expected)
 
 
-def test_resample_groupby_with_label():
+def test_resample_groupby_with_label(unit):
     # GH 13235
-    index = date_range("2000-01-01", freq="2D", periods=5)
+    index = date_range("2000-01-01", freq="2D", periods=5, unit=unit)
     df = DataFrame(index=index, data={"col0": [0, 0, 1, 1, 2], "col1": [1, 1, 1, 1, 1]})
     msg = "DataFrameGroupBy.resample operated on the grouping columns"
     with tm.assert_produces_warning(FutureWarning, match=msg):
         result = df.groupby("col0").resample("1W", label="left").sum()
 
     mi = [
         np.array([0, 0, 1, 2], dtype=np.int64),
-        pd.to_datetime(
-            np.array(["1999-12-26", "2000-01-02", "2000-01-02", "2000-01-02"])
+        np.array(
+            ["1999-12-26", "2000-01-02", "2000-01-02", "2000-01-02"],
+            dtype=f"M8[{unit}]",
         ),
     ]
     mindex = pd.MultiIndex.from_arrays(mi, names=["col0", None])
@@ -505,7 +506,9 @@ def test_resample_groupby_agg_object_dtype_all_nan(consolidate):
     idx = pd.MultiIndex.from_arrays(
         [
             ["A"] * 3 + ["B"] * 3,
-            pd.to_datetime(["2020-01-05", "2020-01-12", "2020-01-19"] * 2),
+            pd.to_datetime(["2020-01-05", "2020-01-12", "2020-01-19"] * 2).as_unit(
+                "ns"
+            ),
         ],
         names=["key", "date"],
     )
@@ -530,19 +533,15 @@ def test_groupby_resample_with_list_of_keys():
         }
     )
     result = df.groupby("group").resample("2D", on="date")[["val"]].mean()
+
+    mi_exp = pd.MultiIndex.from_arrays(
+        [[0, 0, 1, 1], df["date"]._values[::2]], names=["group", "date"]
+    )
     expected = DataFrame(
         data={
             "val": [4.0, 3.5, 6.5, 3.0],
         },
-        index=Index(
-            data=[
-                (0, Timestamp("2016-01-01")),
-                (0, Timestamp("2016-01-03")),
-                (1, Timestamp("2016-01-05")),
-                (1, Timestamp("2016-01-07")),
-            ],
-            name=("group", "date"),
-        ),
+        index=mi_exp,
     )
     tm.assert_frame_equal(result, expected)
 
@@ -605,17 +604,17 @@ def test_groupby_resample_size_all_index_same():
     msg = "DataFrameGroupBy.resample operated on the grouping columns"
     with tm.assert_produces_warning(FutureWarning, match=msg):
         result = df.groupby("A").resample("D").size()
+
+    mi_exp = pd.MultiIndex.from_arrays(
+        [
+            [1, 1, 2, 2],
+            pd.DatetimeIndex(["2000-12-31", "2001-01-01"] * 2, dtype="M8[ns]"),
+        ],
+        names=["A", None],
+    )
     expected = Series(
         3,
-        index=pd.MultiIndex.from_tuples(
-            [
-                (1, Timestamp("2000-12-31")),
-                (1, Timestamp("2001-01-01")),
-                (2, Timestamp("2000-12-31")),
-                (2, Timestamp("2001-01-01")),
-            ],
-            names=["A", None],
-        ),
+        index=mi_exp,
     )
     tm.assert_series_equal(result, expected)
 
@@ -627,25 +626,18 @@ def test_groupby_resample_on_index_with_list_of_keys():
             "group": [0, 0, 0, 0, 1, 1, 1, 1],
             "val": [3, 1, 4, 1, 5, 9, 2, 6],
         },
-        index=Series(
-            date_range(start="2016-01-01", periods=8),
-            name="date",
-        ),
+        index=date_range(start="2016-01-01", periods=8, name="date"),
     )
     result = df.groupby("group").resample("2D")[["val"]].mean()
+
+    mi_exp = pd.MultiIndex.from_arrays(
+        [[0, 0, 1, 1], df.index[::2]], names=["group", "date"]
+    )
     expected = DataFrame(
         data={
             "val": [2.0, 2.5, 7.0, 4.0],
         },
-        index=Index(
-            data=[
-                (0, Timestamp("2016-01-01")),
-                (0, Timestamp("2016-01-03")),
-                (1, Timestamp("2016-01-05")),
-                (1, Timestamp("2016-01-07")),
-            ],
-            name=("group", "date"),
-        ),
+        index=mi_exp,
     )
     tm.assert_frame_equal(result, expected)
 
@@ -659,26 +651,19 @@ def test_groupby_resample_on_index_with_list_of_keys_multi_columns():
             "second_val": [2, 7, 1, 8, 2, 8, 1, 8],
             "third_val": [1, 4, 1, 4, 2, 1, 3, 5],
         },
-        index=Series(
-            date_range(start="2016-01-01", periods=8),
-            name="date",
-        ),
+        index=date_range(start="2016-01-01", periods=8, name="date"),
     )
     result = df.groupby("group").resample("2D")[["first_val", "second_val"]].mean()
+
+    mi_exp = pd.MultiIndex.from_arrays(
+        [[0, 0, 1, 1], df.index[::2]], names=["group", "date"]
+    )
     expected = DataFrame(
         data={
             "first_val": [2.0, 2.5, 7.0, 4.0],
             "second_val": [4.5, 4.5, 5.0, 4.5],
         },
-        index=Index(
-            data=[
-                (0, Timestamp("2016-01-01")),
-                (0, Timestamp("2016-01-03")),
-                (1, Timestamp("2016-01-05")),
-                (1, Timestamp("2016-01-07")),
-            ],
-            name=("group", "date"),
-        ),
+        index=mi_exp,
     )
     tm.assert_frame_equal(result, expected)