size_hint for hash tables does already the right thing, not need for additional checks

Author: realead

pandas/core/frame.py

@@ -5451,15 +5451,13 @@ def duplicated(
         4     True
         dtype: bool
         """
-        from pandas._libs.hashtable import SIZE_HINT_LIMIT, duplicated_int64
+        from pandas._libs.hashtable import duplicated_int64
 
         if self.empty:
             return self._constructor_sliced(dtype=bool)
 
         def f(vals):
-            labels, shape = algorithms.factorize(
-                vals, size_hint=min(len(self), SIZE_HINT_LIMIT)
-            )
+            labels, shape = algorithms.factorize(vals, size_hint=len(self))
             return labels.astype("i8", copy=False), len(shape)
 
         if subset is None:

pandas/core/sorting.py

@@ -593,7 +593,7 @@ def compress_group_index(group_index, sort: bool = True):
     space can be huge, so this function compresses it, by computing offsets
     (comp_ids) into the list of unique labels (obs_group_ids).
     """
-    size_hint = min(len(group_index), hashtable.SIZE_HINT_LIMIT)
+    size_hint = len(group_index)
     table = hashtable.Int64HashTable(size_hint)
 
     group_index = ensure_int64(group_index)