Not sure if it was ever supposed to work with numpy arrays, but it actually does :thinking::

```python

attr = np.array([[3, 4]]) d1 = xr.Dataset({'z': 1}, attrs={'y': attr}) d2 = xr.Dataset({'z': 2}, attrs={'y': attr.copy()}) xr.concat([d1, d2], dim='z', compat='identical') ```

However, it fails if you use DataArray attrs:

```python

attr = xr.DataArray([3, 4], {'x': [1, 2]}, 'x') d1 = xr.Dataset({'z': 1}, attrs={'y': attr}) d2 = xr.Dataset({'z': 2}, attrs={'y': attr.copy()}) xr.concat([d1, d2], dim='z', compat='identical') ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all() ```

Given that the check is simply (a is b) or (a == b), should it try to do something smarter for array-like attrs?

Currently, to easily create a Dataset from an existing numpy recarray (not a DataArray, which is currently bugged anyway with recarrays due to #1434), I couldn't find an easier way than

python df = xr.Dataset.from_dataframe(pd.DataFrame(my_recarray).set_index('foo'))

(which is kind of dumb since it allocates the memory twice)

It would definitely be nice to be able to do just this (perhaps with extra arguments to set index on the fly etc):

python df = xr.Dataset.from_records(my_recarray, ...)

(Apologies if I'm missing something obvious.)

These are two unrelated issues in one really that I've noticed while trying to plot things directly from DataArray objects.

The following works as expected, by converting DataArray to pandas first)

```python

arr.to_series().plot(style='.-') arr.to_series().plot.line(style='.-') ```

Passing Series to pyplot.plot() directly also works and retains index:

```python

plt.plot(arr.to_series(), '.-') ```

Trying to set style directly when plotting from DataArray doesn't work: ```python

arr.plot(style='.-') AttributeError: Unknown property style arr.plot.line(style='.-') AttributeError: Unknown property style ```

Passing DataArray to pyplot.plot() loses index:

```python

plt.plot(arr, '.-')

works but loses coords; same as plot.plot(arr.values, '.-')

```

(xarray version: 0.11.3)

Just wanted to confirm this is expected behaviour: sel() with a date that would yield an empty selection throws an exception (I would naturally expect it to return a zero-length dataarray/dataset instead):

```python

foo = xr.DataArray( np.array([1, 2, 3]), {'t': pd.to_datetime(['2018-01-01', '2018-02-02T01:01', '2018-02-02T02:02'])}, dims=['t'] )

foo.sel(t='2018-01-01').size 1

foo.sel(t='2018-02-02').size 2

foo.sel(t='2018-03-03').size # expected 0?

pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.Int64HashTable.get_item() TypeError: an integer is required

During handling of the above exception, another exception occurred:

pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine._date_check_type() KeyError: '2018-03-03'

During handling of the above exception, another exception occurred:

pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.Int64HashTable.get_item() TypeError: an integer is required

During handling of the above exception, another exception occurred:

pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine._date_check_type()

KeyError: '2018-03-03'

During handling of the above exception, another exception occurred:

pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.Int64HashTable.get_item() pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.Int64HashTable.get_item() KeyError: 1520035200000000000

During handling of the above exception, another exception occurred:

pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() KeyError: Timestamp('2018-03-03 00:00:00')

During handling of the above exception, another exception occurred:

pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.Int64HashTable.get_item() pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.Int64HashTable.get_item() KeyError: 1520035200000000000

During handling of the above exception, another exception occurred:

pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() pandas/_libs/index.pyx in pandas._libs.index.DatetimeEngine.get_loc() KeyError: Timestamp('2018-03-03 00:00:00')

...

```