FIX-modin-project#5091: Handle pd.Grouper objects correctly

Signed-off-by: Karthik Velayutham <[email protected]>

modin/core/storage_formats/pandas/query_compiler.py

@@ -3114,7 +3114,12 @@ def _groupby_internal_columns(self, by, drop):
         else:
             if not isinstance(by, list):
                 by = [by] if by is not None else []
-            internal_by = [o for o in by if hashable(o) and o in self.columns]
+            internal_by = []
+            for o in by:
+                if isinstance(o, pandas.Grouper):
+                    internal_by.append(o.key)
+                elif hashable(o) and o in self.columns:
+                    internal_by.append(o)
             internal_qc = (
                 [self.getitem_column_array(internal_by)] if len(internal_by) else []
             )

modin/pandas/dataframe.py

@@ -482,29 +482,34 @@ def groupby(
                 (
                     (hashable(o) and (o in self))
                     or isinstance(o, Series)
+                    or (isinstance(o, pandas.Grouper) and o.key in self)
                     or (is_list_like(o) and len(o) == len(self.axes[axis]))
                 )
                 for o in by
             ):
-                # We want to split 'by's into those that belongs to the self (internal_by)
-                # and those that doesn't (external_by)
-                internal_by, external_by = [], []
+                has_external = False
+                processed_by = []
 
                 for current_by in by:
-                    if hashable(current_by):
-                        internal_by.append(current_by)
+                    if isinstance(current_by, pandas.Grouper):
+                        processed_by.append(current_by)
+                        has_external = True
+                    elif hashable(current_by):
+                        processed_by.append(current_by)
                     elif isinstance(current_by, Series):
                         if current_by._parent is self:
-                            internal_by.append(current_by.name)
+                            processed_by.append(current_by.name)
                         else:
-                            external_by.append(current_by._query_compiler)
+                            processed_by.append(current_by._query_compiler)
+                            has_external = True
                     else:
-                        external_by.append(current_by)
+                        has_external = True
+                        processed_by.append(current_by)
 
-                by = internal_by + external_by
+                by = processed_by
 
-                if len(external_by) == 0:
-                    by = self[internal_by]._query_compiler
+                if not has_external:
+                    by = self[processed_by]._query_compiler
 
                 drop = True
             else:

modin/pandas/groupby.py

@@ -573,7 +573,13 @@ def _internal_by(self):
         internal_by = tuple()
         if self._drop:
             if is_list_like(self._by):
-                internal_by = tuple(by for by in self._by if isinstance(by, str))
+                internal_by_list = []
+                for by in self._by:
+                    if isinstance(by, str):
+                        internal_by_list.append(by)
+                    elif isinstance(by, pandas.Grouper):
+                        internal_by_list.append(by.key)
+                internal_by = tuple(internal_by_list)
             else:
                 ErrorMessage.catch_bugs_and_request_email(
                     failure_condition=not isinstance(self._by, BaseQueryCompiler),