Pandas.ix[:, "2002"] returns a DataFrame

[dalejung]

import pandas as pd
import numpy as np

ind = pd.date_range(start="2000", freq="D", periods=1000)
df = pd.DataFrame(np.random.randn(len(ind), 5), index=ind, columns=list('ABCDE'))
panel = pd.Panel({'frame_'+c:df for c in list('ABC')})

test1 = panel.ix[:, "2002"]
test1.ndim # 3
type(test1) # pandas.core.frame.DataFrame

test2 = panel.ix[:, "2002":"2002-12-31"]
test2.ndim # 3
type(test2) # pandas.core.panel.Panel

print pd.__version__
#0.11.1.dev-d7fe745

When trying to grab all the data for year 2002, I get back a DataFrame. If I use a range for the major axis, then it returns a Panel.

http://nbviewer.ipython.org/5853887

[jreback]

yep....not tested real well here; nor is some of this implemented for >2 dim, marking as a bug

[cpcloud]

ha! beat u! sorry :P

[cpcloud]

wonder if we were looking at the notebook at the same time...

[cpcloud]

@jreback should be marked as 0.11.1?

[jreback]

you can do a PR if you want, but marked as 0.12; nothing more for 0.11.1 (unless its ready and tested and can be merged right away)....no new issues

[cpcloud]

excellent.