From 30013ee067e7ebc03ce4e0fe6c0d44c44205f3e4 Mon Sep 17 00:00:00 2001
From: Kei <keiwork111@gmail.com>
Date: Sat, 20 Apr 2024 16:38:15 +0800
Subject: [PATCH] Move calculation of observed grouper to when initialising
 groupby

---
 pandas/core/groupby/generic.py |  6 ++--
 pandas/core/groupby/groupby.py | 56 +++++++++++++---------------------
 2 files changed, 26 insertions(+), 36 deletions(-)

diff --git a/pandas/core/groupby/generic.py b/pandas/core/groupby/generic.py
index bbfc35b8e30b7..d298139d72554 100644
--- a/pandas/core/groupby/generic.py
+++ b/pandas/core/groupby/generic.py
@@ -2040,12 +2040,13 @@ def _gotitem(self, key, ndim: int, subset=None):
                 group_keys=self.group_keys,
                 observed=self.observed,
                 dropna=self.dropna,
+                observed_grouper=self.observed_grouper,
+                observed_exclusions=self.observed_exclusions,
             )
         elif ndim == 1:
             if subset is None:
                 subset = self.obj[key]
 
-            orig_obj = self.orig_obj if not self.observed else None
             return SeriesGroupBy(
                 subset,
                 self.keys,
@@ -2058,7 +2059,8 @@ def _gotitem(self, key, ndim: int, subset=None):
                 group_keys=self.group_keys,
                 observed=self.observed,
                 dropna=self.dropna,
-                orig_obj=orig_obj,
+                observed_grouper=self.observed_grouper,
+                observed_exclusions=self.observed_exclusions,
             )
 
         raise AssertionError("invalid ndim for _gotitem")
diff --git a/pandas/core/groupby/groupby.py b/pandas/core/groupby/groupby.py
index 7661902a65579..c34c3a3861255 100644
--- a/pandas/core/groupby/groupby.py
+++ b/pandas/core/groupby/groupby.py
@@ -586,9 +586,10 @@ class BaseGroupBy(PandasObject, SelectionMixin[NDFrameT], GroupByIndexingMixin):
         "keys",
         "level",
         "obj",
-        "orig_obj",
         "observed",
         "sort",
+        "observed_grouper",
+        "observed_exclusions",
     }
 
     _grouper: ops.BaseGrouper
@@ -1107,7 +1108,8 @@ def __init__(
         group_keys: bool = True,
         observed: bool = False,
         dropna: bool = True,
-        orig_obj: NDFrameT | None = None,
+        observed_grouper: ops.BaseGrouper | None = None,
+        observed_exclusions: frozenset[Hashable] | None = None,
     ) -> None:
         self._selection = selection
 
@@ -1119,8 +1121,8 @@ def __init__(
         self.sort = sort
         self.group_keys = group_keys
         self.dropna = dropna
-        self.orig_obj = obj if orig_obj is None else orig_obj
 
+        orig_obj = obj
         if grouper is None:
             grouper, exclusions, obj = get_grouper(
                 obj,
@@ -1136,6 +1138,21 @@ def __init__(
         self._grouper = grouper
         self.exclusions = frozenset(exclusions) if exclusions else frozenset()
 
+        if not observed and observed_grouper is None:
+            observed_grouper, observed_exclusions, _ = get_grouper(
+                orig_obj,
+                self.keys,
+                level=self.level,
+                sort=self.sort,
+                observed=True,
+                dropna=self.dropna,
+            )
+
+        self.observed_grouper = observed_grouper
+        self.observed_exclusions = (
+            frozenset(observed_exclusions) if observed_exclusions else frozenset()
+        )
+
     def __getattr__(self, attr: str):
         if attr in self._internal_names_set:
             return object.__getattribute__(self, attr)
@@ -1887,44 +1904,15 @@ def _transform(self, func, *args, engine=None, engine_kwargs=None, **kwargs):
                     func, *args, engine=engine, engine_kwargs=engine_kwargs, **kwargs
                 )
 
-            grouper, exclusions, obj = get_grouper(
-                self.orig_obj,
-                self.keys,
-                level=self.level,
-                sort=self.sort,
-                observed=True,
-                dropna=self.dropna,
-            )
-            exclusions = frozenset(exclusions) if exclusions else frozenset()
-            obj_has_not_changed = self.orig_obj.equals(self.obj)
-
             with (
                 com.temp_setattr(self, "observed", True),
-                com.temp_setattr(self, "_grouper", grouper),
-                com.temp_setattr(self, "exclusions", exclusions),
-                com.temp_setattr(self, "obj", obj, condition=obj_has_not_changed),
+                com.temp_setattr(self, "_grouper", self.observed_grouper),
+                com.temp_setattr(self, "exclusions", self.observed_exclusions),
             ):
                 return self._reduction_kernel_transform(
                     func, *args, engine=engine, engine_kwargs=engine_kwargs, **kwargs
                 )
 
-            # with com.temp_setattr(self, "as_index", True):
-            #     # GH#49834 - result needs groups in the index for
-            #     # _wrap_transform_fast_result
-            #     if func in ["idxmin", "idxmax"]:
-            #         func = cast(Literal["idxmin", "idxmax"], func)
-            #         result = self._idxmax_idxmin(func, True, *args, **kwargs)
-            #     else:
-            #         if engine is not None:
-            #             kwargs["engine"] = engine
-            #             kwargs["engine_kwargs"] = engine_kwargs
-            #         result = getattr(self, func)(*args, **kwargs)
-
-            #         print("result with observed = False\n", result.to_string())
-            #         r = self._wrap_transform_fast_result(result)
-            #         print("reindexed result", r.to_string())
-            #         return r
-
     @final
     def _reduction_kernel_transform(
         self, func, *args, engine=None, engine_kwargs=None, **kwargs