BUG: crosstab fails with categoricals on master

[bashtage]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

crosstab no longer works correctly with categorical inputs. This is a regression from 1.1.x.

Code Sample, a copy-pastable example

import pandas as pd
import numpy as np

g = np.random.default_rng(840812492384587325982704)
a = pd.Series(g.integers(0,3,size=100)).astype("category")
b = pd.Series(g.integers(0,2,size=100)).astype("category")
pd.crosstab(a,b,margins=True,dropna=False)

Produces

Traceback (most recent call last):
  File "C:\git\pandas\pandas\core\indexes\base.py", line 2965, in get_loc
    return self._engine.get_loc(casted_key)
  File "pandas\_libs\index.pyx", line 70, in pandas._libs.index.IndexEngine.get_loc
    cpdef get_loc(self, object val):
  File "pandas\_libs\index.pyx", line 98, in pandas._libs.index.IndexEngine.get_loc
    self._check_type(val)
  File "pandas\_libs\index_class_helper.pxi", line 93, in pandas._libs.index.Int64Engine._check_type
    raise KeyError(val)
KeyError: 'All'
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "pandas\_libs\index.pyx", line 705, in pandas._libs.index.BaseMultiIndexCodesEngine.get_loc
    indices = [0 if checknull(v) else lev.get_loc(v) + 1
  File "C:\git\pandas\pandas\core\indexes\base.py", line 2963, in get_loc
    casted_key = self._maybe_cast_indexer(key)
  File "C:\git\pandas\pandas\core\indexes\category.py", line 527, in _maybe_cast_indexer
    return self._data._unbox_scalar(key)
  File "C:\git\pandas\pandas\core\arrays\categorical.py", line 1728, in _unbox_scalar
    code = self.categories.get_loc(key)
  File "C:\git\pandas\pandas\core\indexes\base.py", line 2967, in get_loc
    raise KeyError(key) from err
KeyError: 'All'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "C:\git\pandas\pandas\core\generic.py", line 3776, in _set_item
    loc = self._info_axis.get_loc(key)
  File "C:\git\pandas\pandas\core\indexes\multi.py", line 2795, in get_loc
    return self._engine.get_loc(key)
  File "pandas\_libs\index.pyx", line 708, in pandas._libs.index.BaseMultiIndexCodesEngine.get_loc
    raise KeyError(key)
KeyError: ('__dummy__', 'All')
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "C:\Anaconda\lib\site-packages\IPython\core\interactiveshell.py", line 3417, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-3-ba4a70b3d31b>", line 7, in <module>
    pd.crosstab(a,b,margins=True,dropna=False)
  File "C:\git\pandas\pandas\core\reshape\pivot.py", line 623, in crosstab
    **kwargs,
  File "C:\git\pandas\pandas\core\frame.py", line 6972, in pivot_table
    observed=observed,
  File "C:\git\pandas\pandas\core\reshape\pivot.py", line 180, in pivot_table
    fill_value=fill_value,
  File "C:\git\pandas\pandas\core\reshape\pivot.py", line 242, in _add_margins
    table, data, values, rows, cols, aggfunc, observed, margins_name
  File "C:\git\pandas\pandas\core\reshape\pivot.py", line 331, in _generate_marginal_results
    piece[all_key] = margin[key]
  File "C:\git\pandas\pandas\core\frame.py", line 3083, in __setitem__
    self._set_item(key, value)
  File "C:\git\pandas\pandas\core\frame.py", line 3160, in _set_item
    NDFrame._set_item(self, key, value)
  File "C:\git\pandas\pandas\core\generic.py", line 3779, in _set_item
    self._mgr.insert(len(self._info_axis), key, value)
  File "C:\git\pandas\pandas\core\internals\managers.py", line 1173, in insert
    new_axis = self.items.insert(loc, item)
  File "C:\git\pandas\pandas\core\indexes\multi.py", line 3667, in insert
    level = level.insert(lev_loc, k)
  File "C:\git\pandas\pandas\core\indexes\category.py", line 695, in insert
    code = self._data._validate_insert_value(item)
  File "C:\git\pandas\pandas\core\arrays\categorical.py", line 1186, in _validate_insert_value
    return self._validate_fill_value(value)
  File "C:\git\pandas\pandas\core\arrays\categorical.py", line 1222, in _validate_fill_value
    f"'fill_value={fill_value}' is not present "
ValueError: 'fill_value=All' is not present in this Categorical's categories

Problem description

Previously produced the correct output. A simple regression.

Expected Output

Output of pd.show_versions()

INSTALLED VERSIONS

commit : d89331b
python : 3.7.9.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19041
machine : AMD64
processor : Intel64 Family 6 Model 94 Stepping 3, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : None.None
pandas : 1.2.0.dev0+948.gd89331b96
numpy : 1.19.2
pytz : 2020.1
dateutil : 2.8.1
pip : 20.2.4
setuptools : 50.3.0.post20201006
Cython : 0.29.21
pytest : 6.1.1
hypothesis : None
sphinx : 3.2.1
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 4.4.0
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.11.2
IPython : 7.18.1
pandas_datareader: v0.10.0dev0+30.gadb67b2
bs4 : 4.8.0
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.3.2
numexpr : 2.7.1
odfpy : None
openpyxl : 3.0.5
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.5.2
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : 1.2.0
xlwt : 1.3.0
numba : 0.51.2

[jbrockmendel]

The exception message makes me think this is likely due to one of my recent PRs. I'll take a look.

[bashtage]

I suspect the wrong exception might be being raised. This said, ValueError does seem right when looking up a value in a Categorical index that is missing.

[simonjayhawkins]

This is working on 1.1.3. will bisect shortly.

[simonjayhawkins]

first bad commit: [e2785f7] REF: de-duplicate Categorical validators (#36558)

[bashtage]

I checked the 1.1.x branch and this has not been backported to it.

[bashtage]

Pretty obvious the error type was changed.