ERR: better error reporting for passing list to level grouper on non-MultiIndex

[eXcuvator]

groupy(level=[0]) doesn't work on a unique index. I guess it is semantics whether one allows not only level=0, but also level=[0], but if not, the error message should be different:

df = pd.DataFrame(np.arange(0,9).reshape(3,3))
df.groupby(level=[0])
ValueError: level > 0 only valid with MultiIndex

[ron819]

Any chance this will be fixed soon?

[jreback]

a community pull request would make this happen

[mroeschke]

This looks to work on master, as the operation is okay and shouldn't raise and error I suppose. Could use a test.

In [72]: df = pd.DataFrame(np.arange(0,9).reshape(3,3))
    ...: df.groupby(level=[0]).mean()
Out[72]:
   0  1  2
0  0  1  2
1  3  4  5
2  6  7  8

In [73]: df = pd.DataFrame(np.arange(0,9).reshape(3,3))
    ...: df.groupby(level=0).mean()
Out[73]:
   0  1  2
0  0  1  2
1  3  4  5
2  6  7  8

In [74]: pd.__version__
Out[74]: '0.26.0.dev0+682.g08ab156eb'