add fix for opening zero observation dta files

[bquistorff]

Opening a Stata dta file with no observations (but having variables) resulted in an error. Example file: https://dl.dropboxusercontent.com/u/6705315/no_obs_v115.dta.

>>> import pandas as pd
>>> pd.read_stata("no_obs_v115.dta")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Python27\lib\site-packages\pandas\io\stata.py", line 49, in read_stata
    return reader.data(convert_dates, convert_categoricals, index)
  File "C:\Python27\lib\site-packages\pandas\io\stata.py", line 855, in data
    data = DataFrame(data, columns=self.varlist, index=index)
  File "C:\Python27\lib\site-packages\pandas\core\frame.py", line 255, in __init__
    copy=copy)
  File "C:\Python27\lib\site-packages\pandas\core\frame.py", line 367, in _init_ndarray
    return create_block_manager_from_blocks([values.T], [columns, index])
  File "C:\Python27\lib\site-packages\pandas\core\internals.py", line 3185, in create_block_manager_from_blocks
    construction_error(tot_items, blocks[0].shape[1:], axes, e)
  File "C:\Python27\lib\site-packages\pandas\core\internals.py", line 3166, in construction_error
    passed,implied))
ValueError: Shape of passed values is (0, 0), indices imply (1, 0)
>>> pd.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.6.final.0
python-bits: 32
OS: Windows
OS-release: 7
machine: AMD64
processor: AMD64 Family 16 Model 6 Stepping 3, AuthenticAMD
byteorder: little
LC_ALL: None
LANG: None

pandas: 0.14.0
nose: 1.3.3
Cython: 0.20.1
numpy: 1.8.1
scipy: None
statsmodels: None
IPython: None
sphinx: None
patsy: None
scikits.timeseries: None
dateutil: 2.2
pytz: 2014.3
bottleneck: 0.8.0
tables: None
numexpr: None
matplotlib: None
openpyxl: 1.8.6
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
bq: None
apiclient: None
rpy2: None
sqlalchemy: None
pymysql: None
psycopg2: None

The PR fixes this issue in stata.py, though maybe the issue should be fixed in DataFrame.

[jreback]

needs a test for doing this (e.g. write an empty frame then read it back in and compare), put in io/tests/test_stat.py

[jreback]

you can also add your test file to the repo (same place as existiing test files)

[bquistorff]

OK, I've added a test. I've left out my file, because generating the empty file seems to work just fine. Let me know if something isn't right as I'm new to git.

[jreback]

add the comment GH 7369 here

[jreback]

add a note in Bug Fixes in v0.14.1.txt otherwise looks good

[bquistorff]

OK. Thanks for the help.

[jreback]

ok looks fine

pls squash down to a single commit

[bquistorff]

OK. I think that worked.

[hayd]

@bquistorff There's a merge conflict (most likely in the release notes), do you mind rebasing?

@jreback perhaps we could put like fifty blank lines in each subsection of the release notes and let people insert their's in randomly (to minimise merge conflicts early on in the release cycle) ?

[jreback]

@hayd good idea!
go for it

[hayd]

@jreback done 8633e6f (I bet this'll cause at lease 3 merge conflicts). Is the issue that people appends there release note at the end rather than insert it somewhere randomly? Hmmm, we'll see of this helps at all.

@bquistorff pandas git tip: insert release notes somewhere randomly rather than append at the end! :)

[jreback]

@bquistorff thanks!