to_msgpack() fails when data has "period"

[amanhanda]

to_msgpack() fails when the data has "period". This works in 0.16.2

ISSUE

import pandas as pd
pd.DataFrame({'C': pd.period_range('2015-01-01', periods=3)}).to_msgpack()
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-65-8e5ad9cd6036> in <module>()
----> 1 pd.DataFrame({'C': pd.period_range('2015-01-01', periods=3)}).to_msgpack()

/auto/energymdl2/anaconda/envs/commod_20160516_pd18/lib/python2.7/site-packages/pandas/core/generic.pyc in to_msgpack(self, path_or_buf, encoding, **kwargs)
   1120         from pandas.io import packers
   1121         return packers.to_msgpack(path_or_buf, self, encoding=encoding,
-> 1122                                   **kwargs)
   1123
   1124     def to_sql(self, name, con, flavor='sqlite', schema=None, if_exists='fail',

/auto/energymdl2/anaconda/envs/commod_20160516_pd18/lib/python2.7/site-packages/pandas/io/packers.pyc in to_msgpack(path_or_buf, *args, **kwargs)
    152     elif path_or_buf is None:
    153         buf = compat.BytesIO()
--> 154         writer(buf)
    155         return buf.getvalue()
    156     else:

/auto/energymdl2/anaconda/envs/commod_20160516_pd18/lib/python2.7/site-packages/pandas/io/packers.pyc in writer(fh)
    145     def writer(fh):
    146         for a in args:
--> 147             fh.write(pack(a, **kwargs))
    148
    149     if isinstance(path_or_buf, compat.string_types):

/auto/energymdl2/anaconda/envs/commod_20160516_pd18/lib/python2.7/site-packages/pandas/io/packers.pyc in pack(o, default, encoding, unicode_errors, use_single_float, autoreset, use_bin_type)
    699                   use_single_float=use_single_float,
    700                   autoreset=autoreset,
--> 701                   use_bin_type=use_bin_type).pack(o)
    702
    703

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer.pack (pandas/msgpack/_packer.cpp:3434)()

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer.pack (pandas/msgpack/_packer.cpp:3276)()

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer._pack (pandas/msgpack/_packer.cpp:2433)()

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer._pack (pandas/msgpack/_packer.cpp:3006)()

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer._pack (pandas/msgpack/_packer.cpp:2433)()

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer._pack (pandas/msgpack/_packer.cpp:3006)()

pandas/msgpack/_packer.pyx in pandas.msgpack._packer.Packer._pack (pandas/msgpack/_packer.cpp:3088)()

/auto/energymdl2/anaconda/envs/commod_20160516_pd18/lib/python2.7/site-packages/pandas/io/packers.pyc in encode(obj)
    510         return {u'typ': u'period',
    511                 u'ordinal': obj.ordinal,
--> 512                 u'freq': u(obj.freq)}
    513     elif isinstance(obj, BlockIndex):
    514         return {u'typ': u'block_index',

/auto/energymdl2/anaconda/envs/commod_20160516_pd18/lib/python2.7/site-packages/pandas/compat/__init__.pyc in u(s)
    262
    263     def u(s):
--> 264         return unicode(s, "unicode_escape")
    265
    266     def u_safe(s):

TypeError: coercing to Unicode: need string or buffer, Day found

Expected Output

Works on 0.16.2

In [30]: pd.DataFrame({'C': pd.period_range('2015-01-01', periods=3)}).to_msgpack()
Out[30]: '\x84\xa6blocks\x91\x86\xa5items\x86\xa4name\xc0\xa5dtype\x11\xa8compress\xc0\xa4data\x91\xa1C\xa5klass\xa5Index\xa3typ\xa5index\xa8compress\xc0\xa5shape\x92\x01\x03\xa6values\x93\x83\xa7ordinal\xcd@4\xa4freq\xa1D\xa3typ\xa6period\x83\xa7ordinal\xcd@5\xa4freq\xa1D\xa3typ\xa6period\x83\xa7ordinal\xcd@6\xa4freq\xa1D\xa3typ\xa6period\xa5klass\xabObjectBlock\xa5dtype\x11\xa4axes\x92\x86\xa4name\xc0\xa5dtype\x11\xa8compress\xc0\xa4data\x91\xa1C\xa5klass\xa5Index\xa3typ\xa5index\x86\xa4name\xc0\xa5dtype\x07\xa8compress\xc0\xa4data\xb8\x00\x00\x00\x00\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x02\x00\x00\x00\x00\x00\x00\x00\xa5klass\xaaInt64Index\xa3typ\xa5index\xa3typ\xadblock_manager\xa5klass\xa9DataFrame'

INSTALLED VERSIONS

------------------
commit: None
python: 2.7.11.final.0
python-bits: 64
OS: Linux
OS-release: 2.6.32-573.7.1.el6.x86_64
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: C
LANG: en_US.UTF-8

pandas: 0.18.1
nose: 1.3.7
pip: 8.1.1
setuptools: 20.7.0
Cython: 0.24
numpy: 1.10.4
scipy: 0.17.0
statsmodels: 0.6.1
xarray: 0.7.2
IPython: 4.1.2
sphinx: 1.3.5
patsy: 0.4.1
dateutil: 2.5.2
pytz: 2016.4
blosc: None
bottleneck: 1.0.0
tables: 3.2.2
numexpr: 2.5.2
matplotlib: 1.4.3
openpyxl: 2.3.2
xlrd: 0.9.4
xlwt: 1.0.0
xlsxwriter: 0.8.4
lxml: 3.6.0
bs4: 4.3.2
html5lib: 0.999
httplib2: 0.9.2
apiclient: 1.5.0
sqlalchemy: 1.0.12
pymysql: None
psycopg2: None
jinja2: 2.8
boto: 2.39.0
pandas_datareader: None

[jreback]

So periods work as part of a Period index, but general support for object that is non-string is not very well covered in tests. Currently Period is an object dtype.

[jreback]

quickest way to have this fixed is to do a patch and submit as a PR.

[amanhanda]

For Period, we would patch with obj.freqstr, instead of obj.freq.

    509     elif isinstance(obj, Period):
    510         return {u'typ': u'period',
    511                 u'ordinal': obj.ordinal,
1-> 512                 u'freq': u(obj.freq)}

[jreback]

yes this is prob a simple fix. But needs some adequate testing (as well as prob a checkin of some generated msgpacks on older versions, use pandas/io/tests/generate_legacy_storage_files.py to generate from a working version)