fix bug for default level behavior

Author: ssf0409

pandas/core/generic.py

@@ -4126,7 +4126,11 @@ class   animal   locomotion
             index = self.index
 
         if isinstance(index, MultiIndex):
-            loc, new_index = index._get_loc_level(key, level=0)
+            loc, new_index = index.get_loc_level(
+                key,
+                level=range(len(key)),
+                drop_level=drop_level
+            )
             if not drop_level:
                 if lib.is_integer(loc):
                     # Slice index must be an integer or None

pandas/tests/series/indexing/test_xs.py

@@ -1,7 +1,9 @@
 import numpy as np
 import pytest
+import debugpy
 
 from pandas import (
+    DataFrame,
     MultiIndex,
     Series,
     date_range,
@@ -80,3 +82,16 @@ def test_xs_key_as_list(self):
 
         with pytest.raises(TypeError, match="list keys are not supported"):
             ser.xs(["a"], axis=0, drop_level=False)
+    
+    def test_xs_default_level(self):
+        # GH#59098
+        df = DataFrame(dict(i=[1,2,3], j=[1,1,2], x=[10, 100, 1000])).set_index(["i", "j"])
+        key = (1, 1)
+
+        # Both scenarios should return DataFrame
+        result_with_level = df.xs(key, drop_level=False, level=list(range(len(key))))
+        result_with_default = df.xs(key, drop_level=False)
+
+        assert type(result_with_default) == DataFrame
+        assert type(result_with_level) == type(result_with_default)
+        tm.assert_equal(result_with_level, result_with_default)