PandasArray does not support slicing consistent with other array types

[maartenb]

Code Sample

>>> import pandas as pd
>>> x = pd.Series([1, 2, 3], index=pd.to_datetime(['2020-01-01', '2020-01-02', '2020-01-03']))
>>> x[:,None].ndim
>>> x.to_numpy()[:,None].ndim
>>> x.index.array[:,None].ndim
>>> x.array[:,None].ndim

Problem description

Currently PandasArray behaves differently then either the Series or or np.ndarray.
This causes a problem when passing the PandasArray to matplotlib for example.
PandasArray should implement this consistently like the other cases.

The output from above is:

>>> import pandas as pd
>>> x = pd.Series([1, 2, 3], index=pd.to_datetime(['2020-01-01', '2020-01-02', '2020-01-03']))
>>> x[:,None].ndim
2
>>> x.to_numpy()[:,None].ndim
2
>>> x.index.array[:,None].ndim
2
>>> x.array[:,None].ndim
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/maartenb/anaconda3/envs/Corona/lib/python3.8/site-packages/pandas/core/arrays/numpy_.py", line 240, in __getitem__
    result = type(self)(result)
  File "/Users/maartenb/anaconda3/envs/Corona/lib/python3.8/site-packages/pandas/core/arrays/numpy_.py", line 147, in __init__
    raise ValueError("PandasArray must be 1-dimensional.")
ValueError: PandasArray must be 1-dimensional.

Expected Output

2
2
2
2

Output of pd.show_versions()

Python 3.8.1 (default, Jan  8 2020, 16:15:59) 
[Clang 4.0.1 (tags/RELEASE_401/final)] :: Anaconda, Inc. on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import pandas as pd
>>> pd.show_versions()

INSTALLED VERSIONS

commit : None
python : 3.8.1.final.0
python-bits : 64
OS : Darwin
OS-release : 18.7.0
machine : x86_64
processor : i386
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.0.1
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 46.0.0.post20200309
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : 3.1.3
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
pytest : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None
numba : None

[MillanSharma]

take

[TomAugspurger]

This is intentional. Extension arrays are 1D, so slicing with a 2D indexer is invalid.

[maartenb]

Well, DatetimeArray, IntegerArray are ExtensionArrays too and do accept it.

I get that from a modeling perspective it makes sense, but it breaks matplotlib and possibly other packages relying on this behavior. I guess this originated from numpy.

[jorisvandenbossche]

On Index it is being deprecated (with your example above):

In [9]: x.index[:,None]    
/home/joris/miniconda3/envs/dev/bin/ipython:1: DeprecationWarning: Support for multi-dimensional indexing (e.g. `index[:, None]`) on an Index is deprecated and will be removed in a future version.  Convert to a numpy array before indexing instead.
  #!/home/joris/miniconda3/envs/dev/bin/python
Out[9]: 
array([['2020-01-01T00:00:00.000000000'],
       ['2020-01-02T00:00:00.000000000'],
       ['2020-01-03T00:00:00.000000000']], dtype='datetime64[ns]')

and the plan is to do the same for the other objects. Series is planned, see #27837, but we might have overlooked the extension arrays, as we should probably do the same for those as well

[jbrockmendel]

This now works correctly, closing.