Can't use iloc to set a subset of a categorical dataframe to scalar None.

[mvashishtha]

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): MacOS Big Sur 11.5.2
• Modin version (modin.__version__): 0.13.0+11.g62179ef6
• Python version: 3.9.9
• Code we can use to reproduce:

import modin.pandas as pd
df = pd.DataFrame([["A"]], dtype="category")
df.iloc[0, 0] = None

Describe the problem

I get an assertion error from an internal Modin assertion instead of setting the only element in the dataframe to numpy.NaN, as pandas would.

Source code / logs

Show stack trace

AssertionError                            Traceback (most recent call last)
Input In [20], in <module>
      1 import modin.pandas as pd
      3 df = pd.DataFrame([["A"]], dtype="category")
----> 4 df.iloc[0, 0] = None

File ~/modin/modin/pandas/indexing.py:945, in _iLocIndexer.__setitem__(self, key, item)
    942 self._check_dtypes(col_loc)
    944 row_lookup, col_lookup = self._compute_lookup(row_loc, col_loc)
--> 945 super(_iLocIndexer, self).__setitem__(
    946     row_lookup,
    947     col_lookup,
    948     item,
    949     axis=self._determine_setitem_axis(
    950         row_lookup, col_lookup, row_scalar, col_scalar
    951     ),
    952 )

File ~/modin/modin/pandas/indexing.py:413, in _LocationIndexerBase.__setitem__(self, row_lookup, col_lookup, item, axis)
    411 if not assigning_to_single_category_column:
    412     item = self._broadcast_item(row_lookup, col_lookup, item, to_shape)
--> 413 self._write_items(row_lookup, col_lookup, item)

File ~/modin/modin/pandas/indexing.py:493, in _LocationIndexerBase._write_items(self, row_lookup, col_lookup, item)
    480 def _write_items(self, row_lookup, col_lookup, item):
    481     """
    482     Perform remote write and replace blocks.
    483
   (...)
    491         The new item value that needs to be assigned to `self`.
    492     """
--> 493     new_qc = self.qc.write_items(row_lookup, col_lookup, item)
    494     self.df._create_or_update_from_compiler(new_qc, inplace=True)

File ~/modin/modin/core/storage_formats/pandas/query_compiler.py:3062, in PandasQueryCompiler.write_items(self, row_numeric_index, col_numeric_index, broadcasted_items)
   3059     partition.iloc[row_internal_indices, col_internal_indices] = item
   3060     return partition
-> 3062 new_modin_frame = self._modin_frame.apply_select_indices(
   3063     axis=None,
   3064     func=iloc_mut,
   3065     row_labels=row_numeric_index,
   3066     col_labels=col_numeric_index,
   3067     new_index=self.index,
   3068     new_columns=self.columns,
   3069     keep_remaining=True,
   3070     item_to_distribute=broadcasted_items,
   3071 )
   3072 return self.__constructor__(new_modin_frame)

File ~/modin/modin/core/dataframe/pandas/dataframe/dataframe.py:110, in lazy_metadata_decorator.<locals>.decorator.<locals>.run_f_on_minimally_updated_metadata(self, *args, **kwargs)
    108     elif apply_axis == "rows":
    109         obj._propagate_index_objs(axis=0)
--> 110 result = f(self, *args, **kwargs)
    111 if apply_axis is None and not transpose:
    112     result._deferred_index = self._deferred_index

File ~/modin/modin/core/dataframe/pandas/dataframe/dataframe.py:1981, in PandasDataframe.apply_select_indices(self, axis, func, apply_indices, row_labels, col_labels, new_index, new_columns, keep_remaining, item_to_distribute)
   1979 assert row_labels is not None and col_labels is not None
   1980 assert keep_remaining
-> 1981 assert item_to_distribute is not None
   1982 row_partitions_list = self._get_dict_of_block_index(0, row_labels).items()
   1983 col_partitions_list = self._get_dict_of_block_index(1, col_labels).items()

AssertionError: