BUG: raise error for groupby() with invalid tuple key

closes pandas-dev#18798

Author: toobaz

doc/source/whatsnew/v0.22.0.txt

@@ -310,6 +310,7 @@ Groupby/Resample/Rolling
 ^^^^^^^^^^^^^^^^^^^^^^^^
 
 - Bug when grouping by a single column and aggregating with a class like ``list`` or ``tuple`` (:issue:`18079`)
+- Fixed regression in :func:`DataFrame.groupby` which would not emit an error when called with a tuple key not in the index (:issue:`18798`)
 -
 -
 

pandas/core/groupby.py

@@ -2978,7 +2978,8 @@ def is_in_obj(gpr):
 
 def _is_label_like(val):
     return (isinstance(val, compat.string_types) or
-            (val is not None and is_scalar(val)))
+            (val is not None and is_scalar(val)) or
+            isinstance(val, tuple))
 
 
 def _convert_grouper(axis, grouper):

pandas/tests/groupby/test_groupby.py

@@ -2750,7 +2750,6 @@ def test_tuple_warns_unhashable(self):
 
         assert "Interpreting tuple 'by' as a list" in str(w[0].message)
 
-    @pytest.mark.xfail(reason="GH-18798")
     def test_tuple_correct_keyerror(self):
         # https://github.com/pandas-dev/pandas/issues/18798
         df = pd.DataFrame(1, index=range(3),