ceil/floor for DatetimeIndex

docs/cudf/source/api_docs/index_objects.rst

@@ -280,6 +280,8 @@ Time-specific operations
    :toctree: api/
 
    DatetimeIndex.round
+   DatetimeIndex.ceil
+   DatetimeIndex.floor
 
 Conversion
 ~~~~~~~~~~

python/cudf/cudf/core/frame.py

@@ -3673,6 +3673,13 @@ def ceil(self):
         3    5.0
         dtype: float64
         """
+
+        warnings.warn(
+            "Series.ceil and DataFrame.ceil are deprecated and will be \
+                removed in the future",
+            DeprecationWarning,
+        )
+
         return self._unaryop("ceil")
 
     def floor(self):
@@ -3705,6 +3712,13 @@ def floor(self):
         5    3.0
         dtype: float64
         """
+
+        warnings.warn(
+            "Series.ceil and DataFrame.ceil are deprecated and will be \
+                removed in the future",
+            DeprecationWarning,
+        )
+
         return self._unaryop("floor")
 
     def scale(self):

python/cudf/cudf/core/index.py

@@ -1898,6 +1898,68 @@ def _get_dt_field(self, field):
     def is_boolean(self):
         return False
 
+    def ceil(self, field):
+        """
+        Perform ceil operation on the data to the specified freq.
+
+        Parameters
+        ----------
+        field : str
+            One of ["D", "H", "T", "min", "S", "L", "ms", "U", "us", "N"].
+            Must be a fixed frequency like 'S' (second) not 'ME' (month end).
+            See `frequency aliases <https://pandas.pydata.org/docs/\
+                user_guide/timeseries.html#timeseries-offset-aliases>`__
+            for more details on these aliases.
+
+        Returns
+        -------
+        DatetimeIndex
+            Index of the same type for a DatetimeIndex
+
+        Examples
+        --------
+        >>> import cudf
+        >>> gIndex = cudf.DatetimeIndex(["2020-05-31 08:00:00",
+        ... "1999-12-31 18:40:00"])
+        >>> gIndex.ceil("T")
+        DatetimeIndex(['2020-05-31 08:00:00', '1999-12-31 18:40:00'],
+        dtype='datetime64[ns]', freq=None)
+        """
+        out_column = self._values.ceil(field)
+
+        return self.__class__._from_data({self.name: out_column})
+
+    def floor(self, field):
+        """
+        Perform floor operation on the data to the specified freq.
+
+        Parameters
+        ----------
+        field : str
+            One of ["D", "H", "T", "min", "S", "L", "ms", "U", "us", "N"].
+            Must be a fixed frequency like 'S' (second) not 'ME' (month end).
+            See `frequency aliases <https://pandas.pydata.org/docs/\
+                user_guide/timeseries.html#timeseries-offset-aliases>`__
+            for more details on these aliases.
+
+        Returns
+        -------
+        DatetimeIndex
+            Index of the same type for a DatetimeIndex
+
+        Examples
+        --------
+        >>> import cudf
+        >>> gIndex = cudf.DatetimeIndex(["2020-05-31 08:59:59"
+        ... ,"1999-12-31 18:44:59"])
+        >>> gIndex.floor("T")
+        DatetimeIndex(['2020-05-31 08:59:00', '1999-12-31 18:44:00'],
+        dtype='datetime64[ns]', freq=None)
+        """
+        out_column = self._values.floor(field)
+
+        return self.__class__._from_data({self.name: out_column})
+
 
 class TimedeltaIndex(GenericIndex):
     """

python/cudf/cudf/tests/test_index.py

@@ -2470,3 +2470,29 @@ def test_index_type_methods(data, func):
         assert_eq(False, actual)
     else:
         assert_eq(expected, actual)
+
+
+@pytest.mark.parametrize(
+    "resolution", ["D", "H", "T", "min", "S", "L", "ms", "U", "us", "N"]
+)
+def test_index_datetime_ceil(resolution):
+    cuidx = cudf.DatetimeIndex([1000000, 2000000, 3000000, 4000000, 5000000])
+    pidx = cuidx.to_pandas()
+
+    pidx_ceil = pidx.ceil(resolution)
+    cuidx_ceil = cuidx.ceil(resolution)
+
+    assert_eq(pidx_ceil, cuidx_ceil)
+
+
+@pytest.mark.parametrize(
+    "resolution", ["D", "H", "T", "min", "S", "L", "ms", "U", "us", "N"]
+)
+def test_index_datetime_floor(resolution):
+    cuidx = cudf.DatetimeIndex([1000000, 2000000, 3000000, 4000000, 5000000])
+    pidx = cuidx.to_pandas()
+
+    pidx_floor = pidx.floor(resolution)
+    cuidx_floor = cuidx.floor(resolution)
+
+    assert_eq(pidx_floor, cuidx_floor)