PERF-#5613: Optimize duplicated in case there is only one column partition

modin/core/dataframe/pandas/dataframe/dataframe.py

@@ -1609,7 +1609,7 @@ def _tree_reduce_func(df, *args, **kwargs):
 
         return _tree_reduce_func
 
-    def _compute_tree_reduce_metadata(self, axis, new_parts):
+    def _compute_tree_reduce_metadata(self, axis, new_parts, dtypes=None):
         """
         Compute the metadata for the result of reduce function.
 
@@ -1619,6 +1619,10 @@ def _compute_tree_reduce_metadata(self, axis, new_parts):
             The axis on which reduce function was applied.
         new_parts : NumPy 2D array
             Partitions with the result of applied function.
+        dtypes : str, optional
+            The data types for the result. This is an optimization
+            because there are functions that always result in a particular data
+            type, and this allows us to avoid (re)computing it.
 
         Returns
         -------
@@ -1633,12 +1637,18 @@ def _compute_tree_reduce_metadata(self, axis, new_parts):
         new_axes_lengths[axis] = [1]
         new_axes_lengths[axis ^ 1] = self._axes_lengths[axis ^ 1]
 
-        new_dtypes = None
+        if dtypes == "copy":
+            dtypes = self._dtypes
+        elif dtypes is not None:
+            dtypes = pandas.Series(
+                [np.dtype(dtypes)] * len(new_axes[1]), index=new_axes[1]
+            )
+
         result = self.__constructor__(
             new_parts,
             *new_axes,
             *new_axes_lengths,
-            new_dtypes,
+            dtypes,
         )
         return result
 
@@ -1677,7 +1687,7 @@ def reduce(
         new_parts = self._partition_mgr_cls.map_axis_partitions(
             axis.value, self._partitions, function
         )
-        return self._compute_tree_reduce_metadata(axis.value, new_parts)
+        return self._compute_tree_reduce_metadata(axis.value, new_parts, dtypes=dtypes)
 
     @lazy_metadata_decorator(apply_axis="opposite", axis_arg=0)
     def tree_reduce(
@@ -2767,24 +2777,27 @@ def broadcast_apply_full_axis(
                     kw["row_lengths"] = get_length_list(
                         axis_len=len(new_index), num_splits=new_partitions.shape[0]
                     )
-                elif (
-                    axis == 1
-                    and self._row_lengths_cache is not None
-                    and len(new_index) == sum(self._row_lengths_cache)
-                ):
-                    kw["row_lengths"] = self._row_lengths_cache
+                elif axis == 1:
+                    if self._row_lengths_cache is not None and len(new_index) == sum(
+                        self._row_lengths_cache
+                    ):
+                        kw["row_lengths"] = self._row_lengths_cache
+                    elif len(new_index) == 1:
+                        kw["row_lengths"] = [1]
             if kw["column_widths"] is None and new_columns is not None:
                 if axis == 1:
                     kw["column_widths"] = get_length_list(
                         axis_len=len(new_columns),
                         num_splits=new_partitions.shape[1],
                     )
-                elif (
-                    axis == 0
-                    and self._column_widths_cache is not None
-                    and len(new_columns) == sum(self._column_widths_cache)
-                ):
-                    kw["column_widths"] = self._column_widths_cache
+                elif axis == 0:
+                    if self._column_widths_cache is not None and len(
+                        new_columns
+                    ) == sum(self._column_widths_cache):
+                        kw["column_widths"] = self._column_widths_cache
+                    elif len(new_columns) == 1:
+                        kw["column_widths"] = [1]
+
         result = self.__constructor__(
             new_partitions, index=new_index, columns=new_columns, **kw
         )

modin/core/storage_formats/pandas/query_compiler.py

@@ -2476,35 +2476,44 @@ def drop(self, index=None, columns=None, errors: str = "raise"):
 
     def duplicated(self, **kwargs):
         def _compute_hash(df):
-            return df.apply(
+            result = df.apply(
                 lambda s: hashlib.new("md5", str(tuple(s)).encode()).hexdigest(), axis=1
-            ).to_frame()
+            )
+            if isinstance(result, pandas.Series):
+                result = result.to_frame(
+                    result.name
+                    if result.name is not None
+                    else MODIN_UNNAMED_SERIES_LABEL
+                )
+            return result
 
         def _compute_duplicated(df):
-            return df.duplicated(**kwargs).to_frame()
+            result = df.duplicated(**kwargs)
+            if isinstance(result, pandas.Series):
+                result = result.to_frame(
+                    result.name
+                    if result.name is not None
+                    else MODIN_UNNAMED_SERIES_LABEL
+                )
+            return result
 
-        new_index = self._modin_frame._index_cache
-        new_columns = [MODIN_UNNAMED_SERIES_LABEL]
-        if len(self.columns) > 1:
-            # if the number of columns we are checking for duplicates is larger than 1,
-            # we must hash them to generate a single value that can be compared across rows.
-            hashed_modin_frame = self._modin_frame.apply_full_axis(
-                1,
-                _compute_hash,
-                new_index=new_index,
-                new_columns=new_columns,
-                keep_partitioning=False,
+        if self._modin_frame._partitions.shape[1] > 1:
+            # if the number of columns (or column partitions) we are checking for duplicates is larger than 1,
+            # we must first hash them to generate a single value that can be compared across rows.
+            hashed_modin_frame = self._modin_frame.reduce(
+                axis=1,
+                function=_compute_hash,
                 dtypes=np.dtype("O"),
             )
         else:
             hashed_modin_frame = self._modin_frame
         new_modin_frame = hashed_modin_frame.apply_full_axis(
-            0,
-            _compute_duplicated,
-            new_index=new_index,
-            new_columns=new_columns,
-            keep_partitioning=False,
+            axis=0,
+            func=_compute_duplicated,
+            new_index=self._modin_frame._index_cache,
+            new_columns=[MODIN_UNNAMED_SERIES_LABEL],
             dtypes=np.bool_,
+            keep_partitioning=False,
         )
         return self.__constructor__(new_modin_frame, shape_hint="column")
 

modin/pandas/dataframe.py

@@ -318,8 +318,6 @@ def duplicated(self, subset=None, keep="first"):  # noqa: PR01, RT01, D200
         df = self[subset] if subset is not None else self
         new_qc = df._query_compiler.duplicated(keep=keep)
         duplicates = self._reduce_dimension(new_qc)
-        # remove Series name which was assigned automatically by .apply in QC
-        duplicates.name = None
         return duplicates
 
     @property