From 6efc2432c062962c0b7c7b9026dd34f42a2ada5f Mon Sep 17 00:00:00 2001
From: Aaron Straus <forbdonut@gmail.com>
Date: Wed, 28 Jun 2017 08:32:30 -0700
Subject: [PATCH] BUG: Fix read of py3 PeriodIndex DataFrame HDF made in py2
 (#16781)

In Python3, reading a DataFrame with a PeriodIndex from an HDF file
created in Python2 would incorrectly return a DataFrame with an
Int64Index.
---
 doc/source/whatsnew/v0.20.3.txt               |   1 +
 pandas/io/pytables.py                         |   4 ++--
 ...periodindex_0.20.1_x86_64_darwin_2.7.13.h5 | Bin 0 -> 7312 bytes
 pandas/tests/io/test_pytables.py              |  20 ++++++++++++++++++
 4 files changed, 23 insertions(+), 2 deletions(-)
 create mode 100644 pandas/tests/io/data/periodindex_0.20.1_x86_64_darwin_2.7.13.h5

diff --git a/doc/source/whatsnew/v0.20.3.txt b/doc/source/whatsnew/v0.20.3.txt
index 0fac6367fd3a5..ebbdecb7dc0d0 100644
--- a/doc/source/whatsnew/v0.20.3.txt
+++ b/doc/source/whatsnew/v0.20.3.txt
@@ -38,6 +38,7 @@ Bug Fixes
 ~~~~~~~~~
 - Fixed issue with dataframe scatter plot for categorical data that reports incorrect column key not found when categorical data is used for plotting (:issue:`16199`)
 - Fixed a pytest marker failing downstream packages' tests suites (:issue:`16680`)
+- Fixed compat with loading a ``DataFrame`` with a ``PeriodIndex``, from a ``format='fixed'`` HDFStore, in Python 3, that was written in Python 2 (:issue:`16781`)
 
 
 
diff --git a/pandas/io/pytables.py b/pandas/io/pytables.py
index 02b6a772e86b9..4e343556c083b 100644
--- a/pandas/io/pytables.py
+++ b/pandas/io/pytables.py
@@ -2591,8 +2591,8 @@ def read_index_node(self, node, start=None, stop=None):
         if 'name' in node._v_attrs:
             name = _ensure_str(node._v_attrs.name)
 
-        index_class = self._alias_to_class(getattr(node._v_attrs,
-                                                   'index_class', ''))
+        index_class = self._alias_to_class(_ensure_decoded(
+            getattr(node._v_attrs, 'index_class', '')))
         factory = self._get_index_factory(index_class)
 
         kwargs = {}
diff --git a/pandas/tests/io/data/periodindex_0.20.1_x86_64_darwin_2.7.13.h5 b/pandas/tests/io/data/periodindex_0.20.1_x86_64_darwin_2.7.13.h5
new file mode 100644
index 0000000000000000000000000000000000000000..6fb92d3c564bdfc290313ab4ef0ca63e1574711e
GIT binary patch
literal 7312
zcmeGgL2}zhY-u|Qvr$r|Z91Jy558HKY9!Zb(^IU}a;+KH@<>X?KG8I__OO;nmP@5^
zUy$DVj2`=h9CPVR9*}l=%rQSmfnC5<Y18fW<iw(_#V!_DEI|MZ?9Ur(>o?_HDe!S!
zu^_6Z%gtMQe)-H|c+B&A)Nu$tBlzo_?KtA9z`0KeU#0aI=Plwj;Rn5TTL2}m&yc3Y
zi>tQhGGYcVl?FE2t!;ea9eQiF(idcBtc-*@Y>yi}@ju~hve%+rS47RMi1zKm-=H(t
z1OpZy!4GA#d_x3Hk^Y;_R=*E;q`v|`rvJ9reS|NX)8p!{c--#wJKZfqkFLY?&3Xey
zNPpiz2Rseww|57vhnwxbx8Ci2*BW?b{LA$w@m0;5ML(^Q!#yW~|K_KE*+_EY?`H8?
zb7?eC^iyk<LL?NwSrPuzP^0A{L<`&>#lca-d*Y8KO6TJ+Q)3*S*!v<5&UpG;-8BcO
z$XtcHV2RjI8Z_Q6?Bk53E;WkJ=jt7+28>1c<nidD#5)`J9M~DgyTp%%zV@=?M49~-
z<LW``k8!3VnD2vWz9%ZxVI0|bj3#noH}&Sd`tmg2b;GBkARdO%p%DfR%EP0G`L^nI
zCYN``>@*sNW1PXbYHfAZ5HcShXF1<3BhUB?xJMeQ>=-yu-Zl6!U#bo#BR_@lO^c%m
zJP+`39E)IjhWx#29~(09+5$L}T(i#H+?*kGVIAzuJGF`>VNbK859w`?=G7;Uc8m5>
zcHXsmz1A-9<n-sCcgyt^Bd{VC4YcR)!m0BY^k`IR<H7>%PKcM+4O?kbrI|gl|1!cF
zP5gGZjKbpiJblic^T{v#?5of17XKjlCqavS^MGDp)k2pbVP{~bftdzo8klKdrh%CT
z-YE^t!#xP*cE!9k&)cq<Jav7u^|;$ReXkKs#tG)F`8~xAlMiR9AL%64YG~DgX(-C!
z?AZ1hH=QS9hU$u=FfyFQd^n6Ub*R#*ywcJ~whSoE*Et_fC|}F#AFdnne7yOe^6^{t
zedxZkn2!@$<nU}gm*?XR%E$kiU|!UG{C%^RyiT)qD43&q>IEZTYh%S|m(%rRqS7!1
zi>-(^X5KNCmMacY^#dZ1{&)DX-3LLO9D7jRNoayN(pfqQvN)AVQ`Q1o_0}^Dt14ZO
z4-T}-w0t<xVWhNP^N(e+ECbKmpM;|<j66@q``@b|lgSryE9)gIvNlZa$^KFlr(=H<
z{-l;<@}=xF8d3vtNp_mpIr?#^Vd5*KzK;ki2~c0lVCZKm3&%=gGgPC@hs`x)t<!*Y
z@E_zfVVQj_iP^t)sBYQ;96q@oT@x1~b0MTp#T(X34yDo6GeS6gFKf%z6Tf15_OpC!
z>)P9k52!|56)QKy%3C5V_q&am-<dCG?*Uk!4Ef%?c<+|S|2rmr)4iX$=bH+r=iF~p
d%=dmtf{O8!<ufchx5~fk-aB=_@xH~=zX5BIWmNzG

literal 0
HcmV?d00001

diff --git a/pandas/tests/io/test_pytables.py b/pandas/tests/io/test_pytables.py
index 676c5a21c38ca..69c92dd775b9a 100644
--- a/pandas/tests/io/test_pytables.py
+++ b/pandas/tests/io/test_pytables.py
@@ -5264,6 +5264,26 @@ def test_fspath(self):
             with pd.HDFStore(path) as store:
                 assert os.fspath(store) == str(path)
 
+    def test_read_py2_hdf_file_in_py3(self):
+        # GH 16781
+
+        # tests reading a PeriodIndex DataFrame written in Python2 in Python3
+
+        # the file was generated in Python 2.7 like so:
+        #
+        # df = pd.DataFrame([1.,2,3], index=pd.PeriodIndex(
+        #              ['2015-01-01', '2015-01-02', '2015-01-05'], freq='B'))
+        # df.to_hdf('periodindex_0.20.1_x86_64_darwin_2.7.13.h5', 'p')
+
+        expected = pd.DataFrame([1., 2, 3], index=pd.PeriodIndex(
+            ['2015-01-01', '2015-01-02', '2015-01-05'], freq='B'))
+
+        with ensure_clean_store(
+                tm.get_data_path('periodindex_0.20.1_x86_64_darwin_2.7.13.h5'),
+                mode='r') as store:
+            result = store['p']
+            assert_frame_equal(result, expected)
+
 
 class TestHDFComplexValues(Base):
     # GH10447