Align misc DataFrame and MultiIndex methods with pandas 2.x (#16402)

The API changes in this PR are mostly adding implementations or adding missing keyword argument (although they might not be implemented). The APIs affected are:

* `DataFrame.insert`
* `DataFrame.melt`
* `DataFrame.merge`
* `DataFrame.quantile`
* `DataFrame.cov`
* `DataFrame.corr`
* `DataFrame.median`
* `DataFrame.rolling`
* `DataFrame.resample`
* `DataFrame.dropna`
* `MultiIndex.from_tuple`
* `MultiIndex.from_frame`
* `MultiIndex.from_product`

Authors:
  - Matthew Roeschke (https://github.com/mroeschke)
  - GALI PREM SAGAR (https://github.com/galipremsagar)

Approvers:
  - GALI PREM SAGAR (https://github.com/galipremsagar)

URL: #16402

python/cudf/cudf/core/dataframe.py

@@ -3215,26 +3215,37 @@ def reset_index(
         )
 
     @_performance_tracking
-    def insert(self, loc, name, value, nan_as_null=no_default):
+    def insert(
+        self,
+        loc,
+        column,
+        value,
+        allow_duplicates: bool = False,
+        nan_as_null=no_default,
+    ):
         """Add a column to DataFrame at the index specified by loc.
 
         Parameters
         ----------
         loc : int
             location to insert by index, cannot be greater then num columns + 1
-        name : number or string
-            name or label of column to be inserted
+        column : number or string
+            column or label of column to be inserted
         value : Series or array-like
         nan_as_null : bool, Default None
             If ``None``/``True``, converts ``np.nan`` values to
             ``null`` values.
             If ``False``, leaves ``np.nan`` values as is.
         """
+        if allow_duplicates is not False:
+            raise NotImplementedError(
+                "allow_duplicates is currently not implemented."
+            )
         if nan_as_null is no_default:
             nan_as_null = not cudf.get_option("mode.pandas_compatible")
         return self._insert(
             loc=loc,
-            name=name,
+            name=column,
             value=value,
             nan_as_null=nan_as_null,
             ignore_index=False,
@@ -4097,7 +4108,15 @@ def transpose(self):
     T = property(transpose, doc=transpose.__doc__)
 
     @_performance_tracking
-    def melt(self, **kwargs):
+    def melt(
+        self,
+        id_vars=None,
+        value_vars=None,
+        var_name=None,
+        value_name="value",
+        col_level=None,
+        ignore_index: bool = True,
+    ):
         """Unpivots a DataFrame from wide format to long format,
         optionally leaving identifier variables set.
 
@@ -4124,23 +4143,30 @@ def melt(self, **kwargs):
         """
         from cudf.core.reshape import melt
 
-        return melt(self, **kwargs)
+        return melt(
+            self,
+            id_vars=id_vars,
+            value_vars=value_vars,
+            var_name=var_name,
+            value_name=value_name,
+            col_level=col_level,
+            ignore_index=ignore_index,
+        )
 
     @_performance_tracking
     def merge(
         self,
         right,
+        how="inner",
         on=None,
         left_on=None,
         right_on=None,
         left_index=False,
         right_index=False,
-        how="inner",
         sort=False,
-        lsuffix=None,
-        rsuffix=None,
-        indicator=False,
         suffixes=("_x", "_y"),
+        indicator=False,
+        validate=None,
     ):
         """Merge GPU DataFrame objects by performing a database-style join
         operation by columns or indexes.
@@ -4241,17 +4267,8 @@ def merge(
             raise NotImplementedError(
                 "Only indicator=False is currently supported"
             )
-
-        if lsuffix or rsuffix:
-            raise ValueError(
-                "The lsuffix and rsuffix keywords have been replaced with the "
-                "``suffixes=`` keyword.  "
-                "Please provide the following instead: \n\n"
-                "    suffixes=('%s', '%s')"
-                % (lsuffix or "_x", rsuffix or "_y")
-            )
-        else:
-            lsuffix, rsuffix = suffixes
+        if validate is not None:
+            raise NotImplementedError("validate is currently not supported.")
 
         lhs, rhs = self, right
         merge_cls = Merge
@@ -5952,9 +5969,9 @@ def quantile(
         axis=0,
         numeric_only=True,
         interpolation=None,
+        method="single",
         columns=None,
         exact=True,
-        method="single",
     ):
         """
         Return values at the given quantile.
@@ -5980,14 +5997,14 @@ def quantile(
                 * higher: `j`.
                 * nearest: `i` or `j` whichever is nearest.
                 * midpoint: (`i` + `j`) / 2.
-        columns : list of str
-            List of column names to include.
-        exact : boolean
-            Whether to use approximate or exact quantile algorithm.
         method : {'single', 'table'}, default `'single'`
             Whether to compute quantiles per-column ('single') or over all
             columns ('table'). When 'table', the only allowed interpolation
             methods are 'nearest', 'lower', and 'higher'.
+        columns : list of str
+            List of column names to include.
+        exact : boolean
+            Whether to use approximate or exact quantile algorithm.
 
         Returns
         -------
@@ -7309,25 +7326,47 @@ def unnamed_group_generator():
             return result
 
     @_performance_tracking
-    def cov(self, **kwargs):
+    def cov(self, min_periods=None, ddof: int = 1, numeric_only: bool = False):
         """Compute the covariance matrix of a DataFrame.
 
         Parameters
         ----------
-        **kwargs
-            Keyword arguments to be passed to cupy.cov
+        min_periods : int, optional
+            Minimum number of observations required per pair of columns to
+            have a valid result.
+            Currently not supported.
+
+        ddof : int, default 1
+            Delta degrees of freedom.  The divisor used in calculations
+            is ``N - ddof``, where ``N`` represents the number of elements.
+
+        numeric_only : bool, default False
+            Include only `float`, `int` or `boolean` data.
+            Currently not supported.
 
         Returns
         -------
         cov : DataFrame
         """
-        cov = cupy.cov(self.values, rowvar=False)
+        if min_periods is not None:
+            raise NotImplementedError(
+                "min_periods is currently not supported."
+            )
+
+        if numeric_only is not False:
+            raise NotImplementedError(
+                "numeric_only is currently not supported."
+            )
+
+        cov = cupy.cov(self.values, ddof=ddof, rowvar=False)
         cols = self._data.to_pandas_index()
         df = DataFrame(cupy.asfortranarray(cov)).set_index(cols)
         df._set_columns_like(self._data)
         return df
 
-    def corr(self, method="pearson", min_periods=None):
+    def corr(
+        self, method="pearson", min_periods=None, numeric_only: bool = False
+    ):
         """Compute the correlation matrix of a DataFrame.
 
         Parameters
@@ -7357,6 +7396,11 @@ def corr(self, method="pearson", min_periods=None):
         if min_periods is not None:
             raise NotImplementedError("Unsupported argument 'min_periods'")
 
+        if numeric_only is not False:
+            raise NotImplementedError(
+                "numeric_only is currently not supported."
+            )
+
         corr = cupy.corrcoef(values, rowvar=False)
         cols = self._data.to_pandas_index()
         df = DataFrame(cupy.asfortranarray(corr)).set_index(cols)

python/cudf/cudf/core/indexed_frame.py

@@ -1495,9 +1495,7 @@ def mean(self, axis=0, skipna=True, numeric_only=False, **kwargs):
             **kwargs,
         )
 
-    def median(
-        self, axis=None, skipna=True, level=None, numeric_only=None, **kwargs
-    ):
+    def median(self, axis=None, skipna=True, numeric_only=None, **kwargs):
         """
         Return the median of the values for the requested axis.
 
@@ -1857,15 +1855,28 @@ def mask(
     @_performance_tracking
     @copy_docstring(Rolling)
     def rolling(
-        self, window, min_periods=None, center=False, axis=0, win_type=None
+        self,
+        window,
+        min_periods=None,
+        center: bool = False,
+        win_type: str | None = None,
+        on=None,
+        axis=0,
+        closed: str | None = None,
+        step: int | None = None,
+        method: str = "single",
     ):
         return Rolling(
             self,
             window,
             min_periods=min_periods,
             center=center,
             axis=axis,
+            on=on,
             win_type=win_type,
+            closed=closed,
+            step=step,
+            method=method,
         )
 
     @copy_docstring(ExponentialMovingWindow)
@@ -1880,6 +1891,7 @@ def ewm(
         ignore_na: bool = False,
         axis: int = 0,
         times: str | np.ndarray | None = None,
+        method: Literal["single", "table"] = "single",
     ):
         return ExponentialMovingWindow(
             self,
@@ -1892,6 +1904,7 @@ def ewm(
             ignore_na=ignore_na,
             axis=axis,
             times=times,
+            method=method,
         )
 
     @_performance_tracking
@@ -3943,16 +3956,15 @@ def resample(
         self,
         rule,
         axis=0,
-        closed=None,
-        label=None,
-        convention="start",
+        closed: Literal["right", "left"] | None = None,
+        label: Literal["right", "left"] | None = None,
+        convention: Literal["start", "end", "s", "e"] = "start",
         kind=None,
-        loffset=None,
-        base=None,
         on=None,
         level=None,
         origin="start_day",
         offset=None,
+        group_keys: bool = False,
     ):
         """
         Convert the frequency of ("resample") the given time series data.
@@ -4090,26 +4102,27 @@ def resample(
                 "deprecated and will be removed in a future version. ",
                 FutureWarning,
             )
-        if (axis, convention, kind, loffset, base, origin, offset) != (
-            0,
-            "start",
-            None,
-            None,
-            None,
-            "start_day",
-            None,
-        ):
-            raise NotImplementedError(
-                "The following arguments are not "
-                "currently supported by resample:\n\n"
-                "- axis\n"
-                "- convention\n"
-                "- kind\n"
-                "- loffset\n"
-                "- base\n"
-                "- origin\n"
-                "- offset"
+            raise NotImplementedError("kind is currently not supported.")
+        if axis != 0:
+            warnings.warn(
+                "The 'axis' keyword in is "
+                "deprecated and will be removed in a future version. ",
+                FutureWarning,
             )
+            raise NotImplementedError("axis is currently not supported.")
+        if convention != "start":
+            warnings.warn(
+                "The 'convention' keyword in is "
+                "deprecated and will be removed in a future version. ",
+                FutureWarning,
+            )
+            raise NotImplementedError("convention is currently not supported.")
+        if origin != "start_day":
+            raise NotImplementedError("origin is currently not supported.")
+        if offset is not None:
+            raise NotImplementedError("offset is currently not supported.")
+        if group_keys is not False:
+            raise NotImplementedError("group_keys is currently not supported.")
         by = cudf.Grouper(
             key=on, freq=rule, closed=closed, label=label, level=level
         )
@@ -4120,7 +4133,13 @@ def resample(
         )
 
     def dropna(
-        self, axis=0, how="any", thresh=None, subset=None, inplace=False
+        self,
+        axis=0,
+        how="any",
+        thresh=None,
+        subset=None,
+        inplace=False,
+        ignore_index: bool = False,
     ):
         """
         Drop rows (or columns) containing nulls from a Column.
@@ -4144,6 +4163,8 @@ def dropna(
             columns, subset is a list of rows to consider.
         inplace : bool, default False
             If True, do operation inplace and return None.
+        ignore_index : bool, default ``False``
+            If ``True``, the resulting axis will be labeled 0, 1, …, n - 1.
 
         Returns
         -------
@@ -4220,6 +4241,8 @@ def dropna(
         """
         if axis == 0:
             result = self._drop_na_rows(how=how, subset=subset, thresh=thresh)
+            if ignore_index:
+                result.index = RangeIndex(len(result))
         else:
             result = self._drop_na_columns(
                 how=how, subset=subset, thresh=thresh