Add condition for test_groupby_nulls_basic in pandas 2.2 (#15072)

This case for some reason doesn't raise a FutureWarning in pandas in 2.2 while it does in pandas 2.1. It's likely a won't-fix so adding a condition

Authors:
  - Matthew Roeschke (https://github.com/mroeschke)

Approvers:
  - Michael Wang (https://github.com/isVoid)

URL: #15072

Author: mroeschke

python/cudf/cudf/tests/test_groupby.py

@@ -1433,7 +1433,7 @@ def test_groupby_nulls_basic(agg):
 
     # TODO: fillna() used here since we don't follow
     # Pandas' null semantics. Should we change it?
-    with expect_warning_if(agg in {"idxmax", "idxmin"}):
+    with expect_warning_if(agg in {"idxmax", "idxmin"} and not PANDAS_GE_220):
         assert_groupby_results_equal(
             getattr(pdf.groupby("a"), agg)().fillna(0),
             getattr(gdf.groupby("a"), agg)().fillna(0 if agg != "prod" else 1),