Slicing a dataframe for rows by set of index labels returns empty rows for missing labels, feature or bug?

[ruidc]

what is the best way to safely slice using multiple index labels? (i.e. not create new NaN rows?) is the best way to do this to intersect the index with the searched values before searching?

import pandas
df = pandas.DataFrame([[1,2], [3,4]], index=['X', 'Y'], columns=['A', 'B'])
df.ix[['X', 'Z'], :]

Out[1]: 
    A   B
X   1   2
Z NaN NaN    < -- why does this not error?

[jreback]

this is essentially a reindex, not intuitive, see #2922 and new docs for .loc, which have different semantics (and won't do this)

[michaelaye]

That's a very meta comment, Jeff. ;) Not sure, I'm getting it...

[jreback]

my comment ? :)

or the issue?

[jreback]

dupe of #2033