BUG: Properly read Categorical msgpacks

Patches bug in read_msgpack in which
Series categoricals were accidentally
being constructed with a non-categorical
dtype, resulting in an error.

Closes pandas-devgh-14901.

Author: gfyoung

doc/source/whatsnew/v0.20.0.txt

@@ -248,6 +248,7 @@ Bug Fixes
 
 
 
+- Bug in ``pd.read_msgpack()`` in which ``Series`` categoricals were being improperly processed (:issue:`14901`)
 
 
 

pandas/io/packers.py

@@ -70,7 +70,7 @@
     move_into_mutable_buffer as _move_into_mutable_buffer,
 )
 
-# check whcih compression libs we have installed
+# check which compression libs we have installed
 try:
     import zlib
 
@@ -591,19 +591,21 @@ def decode(obj):
                           name=obj[u'name'])
 
     elif typ == u'series':
+        name = obj[u'name']
+        index = obj[u'index']
+        klass = globals()[obj[u'klass']]
+
         dtype = dtype_for(obj[u'dtype'])
         pd_dtype = pandas_dtype(dtype)
-        np_dtype = pandas_dtype(dtype).base
-
-        index = obj[u'index']
-        result = globals()[obj[u'klass']](unconvert(obj[u'data'], dtype,
-                                                    obj[u'compress']),
-                                          index=index,
-                                          dtype=np_dtype,
-                                          name=obj[u'name'])
         tz = getattr(pd_dtype, 'tz', None)
+
+        arr = unconvert(obj[u'data'], dtype, obj[u'compress'])
+
         if tz:
+            result = klass(arr, index=index, dtype=pd_dtype.base, name=name)
             result = result.dt.tz_localize('UTC').dt.tz_convert(tz)
+        else:
+            result = klass(arr, index=index, dtype=pd_dtype, name=name)
         return result
 
     elif typ == u'block_manager':

pandas/io/tests/test_packers.py

@@ -363,13 +363,17 @@ def setUp(self):
             'F': [Timestamp('20130102', tz='US/Eastern')] * 2 +
                  [Timestamp('20130603', tz='CET')] * 3,
             'G': [Timestamp('20130102', tz='US/Eastern')] * 5,
+            'H': Categorical([1, 2, 3, 4, 5]),
+            'I': Categorical([1, 2, 3, 4, 5], ordered=True),
         }
 
         self.d['float'] = Series(data['A'])
         self.d['int'] = Series(data['B'])
         self.d['mixed'] = Series(data['E'])
         self.d['dt_tz_mixed'] = Series(data['F'])
         self.d['dt_tz'] = Series(data['G'])
+        self.d['cat_ordered'] = Series(data['H'])
+        self.d['cat_unordered'] = Series(data['I'])
 
     def test_basic(self):