[BUG] In-place updates with loc or iloc don't work correctly when the LHS has more than one column

[magnatelee]

Describe the bug
In-place updates with loc seem to work only when the RHS is broadcasted. The following enumerates three cases that work just fine in Pandas but crash in cuDF. The same issue exists in iloc.

Steps/Code to reproduce bug

(Stacktraces are elided for brevity.)

Bug 1:

>>> x = cudf.DataFrame({"x" : [1,2,3], "y" : [4,5,6]})
>>> x.loc[[0], ["x", "y"]] = [10, 20]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  ...
  File "/home/nfs/wonchanl/anaconda3/envs/rapids-tpcx-bb/lib/python3.7/site-packages/cudf/core/column/column.py", line 693, in __setitem__
    raise ValueError(msg)
ValueError: Size mismatch: cannot set value of size 2 to indexing result of size 1

# The same code works in Pandas
>>> x = pd.DataFrame({"x" : [1,2,3], "y" : [4,5,6]})
>>> x.loc[[0], ["x", "y"]] = [10, 20]
>>> x
    x   y
0  10  20
1   2   5
2   3   6

Bug 2:

>>> x = cudf.DataFrame({"x" : [1,2,3], "y" : [4,5,6]})
>>> x.loc[[0,2], ["x", "y"]] = [[10, 30], [20, 40]]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  ...
  File "/home/nfs/wonchanl/anaconda3/envs/rapids-tpcx-bb/lib/python3.7/site-packages/pandas/core/dtypes/common.py", line 189, in <lambda>
    issubclass(tipo, klasses)
TypeError: issubclass() arg 1 must be a class
>>> x = pd.DataFrame({"x" : [1,2,3], "y" : [4,5,6]})
>>> x.loc[[0,2], ["x", "y"]] = [[10, 30], [20, 40]]
>>> x
    x   y
0  10  30
1   2   5
2  20  40

Bug 3 (note that the index of the RHS is set to [0, 2] to make it align with the LHS):

>>> x = cudf.DataFrame({"x" : [1,2,3], "y" : [4,5,6]})                                                                                                                                                                          
>>> x.loc[[0,2], ["x", "y"]] = cudf.DataFrame({"x" : [10, 20], "y" : [30, 40]}, index=cudf.Index([0, 2]))                                                                                                                       
Traceback (most recent call last):                                                                                                                                                                                              
  File "/home/nfs/wonchanl/anaconda3/envs/rapids-tpcx-bb/lib/python3.7/site-packages/cudf/core/column/column.py", line 1842, in as_column                                                                                       
    memoryview(arbitrary), dtype=dtype, nan_as_null=nan_as_null
TypeError: memoryview: a bytes-like object is required, not 'DataFrame'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/nfs/wonchanl/anaconda3/envs/rapids-tpcx-bb/lib/python3.7/site-packages/cudf/core/column/column.py", line 1871, in as_column
    else nan_as_null,
  ...
  File "/home/nfs/wonchanl/anaconda3/envs/rapids-tpcx-bb/lib/python3.7/site-packages/cudf/core/dataframe.py", line 958, in __arrow_array__
    "Implicit conversion to a host PyArrow Table via __arrow_array__ "
TypeError: Implicit conversion to a host PyArrow Table via __arrow_array__ is not allowed, To explicitly construct a PyArrow Table, consider using .to_arrow()

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  ...
  File "/home/nfs/wonchanl/anaconda3/envs/rapids-tpcx-bb/lib/python3.7/site-packages/cudf/core/dataframe.py", line 950, in __array__
    "Implicit conversion to a host NumPy array via __array__ is not "
TypeError: Implicit conversion to a host NumPy array via __array__ is not allowed, To explicitly construct a GPU matrix, consider using .as_gpu_matrix()
To explicitly construct a host matrix, consider using .as_matrix()

>>> x = pd.DataFrame({"x" : [1,2,3], "y" : [4,5,6]})
>>> x.loc[[0,2], ["x", "y"]] = pd.DataFrame({"x" : [10, 20], "y" : [30, 40]}, index=pd.Index([0, 2]))
>>> x
    x   y
0  10  30
1   2   5
2  20  40

Expected behavior
The examples above should work rather than crashing.

Environment overview (please complete the following information)

• Environment location: Bare-metal
• Method of cuDF install: conda (cudf-0.18.0a+157.g04aa30cf11)

[github-actions]

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.