From 8fbd2f3f643d0ebdf27376a9dba5d4bcaee11ee6 Mon Sep 17 00:00:00 2001
From: Andreas Berneryd <andreas.berneryd@gmail.com>
Date: Sat, 19 Aug 2017 16:54:38 +0200
Subject: [PATCH] BUG: Fix strange behaviour of Series.iloc on MultiIndex
 Series (#17148)

---
 doc/source/whatsnew/v0.21.0.txt    | 1 +
 pandas/core/indexing.py            | 3 ++-
 pandas/tests/indexing/test_iloc.py | 8 ++++++++
 3 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/doc/source/whatsnew/v0.21.0.txt b/doc/source/whatsnew/v0.21.0.txt
index 85685ed7b430d9..4be33a6ce4b697 100644
--- a/doc/source/whatsnew/v0.21.0.txt
+++ b/doc/source/whatsnew/v0.21.0.txt
@@ -329,6 +329,7 @@ Indexing
 - Fixes ``DataFrame.loc`` for setting with alignment and tz-aware ``DatetimeIndex`` (:issue:`16889`)
 - Avoids ``IndexError`` when passing an Index or Series to ``.iloc`` with older numpy (:issue:`17193`)
 - Allow unicode empty strings as placeholders in multilevel columns in Python 2 (:issue:`17099`)
+- Bug ``Series`` with ``MultiIndex`` gives surprising ``.iloc`` results when an integer indexer is used (:issue: `17148`)
 
 I/O
 ^^^
diff --git a/pandas/core/indexing.py b/pandas/core/indexing.py
index 109183827de4e8..eda438737ec26a 100755
--- a/pandas/core/indexing.py
+++ b/pandas/core/indexing.py
@@ -146,7 +146,8 @@ def _get_setitem_indexer(self, key):
             return self._convert_tuple(key, is_setter=True)
 
         axis = self.obj._get_axis(0)
-        if isinstance(axis, MultiIndex):
+
+        if isinstance(axis, MultiIndex) and not isinstance(key, int):
             try:
                 return axis.get_loc(key)
             except Exception:
diff --git a/pandas/tests/indexing/test_iloc.py b/pandas/tests/indexing/test_iloc.py
index 31fee303a41e20..47360ca14401e0 100644
--- a/pandas/tests/indexing/test_iloc.py
+++ b/pandas/tests/indexing/test_iloc.py
@@ -269,6 +269,14 @@ def test_iloc_setitem(self):
         expected = Series([0, 1, 0], index=[4, 5, 6])
         tm.assert_series_equal(s, expected)
 
+        df = pd.DataFrame(data=[[1, 2, 5], [1, 3, 6]], columns=['i', 'j', 'k'])
+        df.set_index(['i', 'j'], inplace=True)
+        s = df.k.copy()
+        s.iloc[1] = 1
+        df.values[1][0] = 1
+        expected = df.k
+        tm.assert_series_equal(s, expected)
+
     def test_iloc_setitem_list(self):
 
         # setitem with an iloc list