BUG: ValueError: buffer source array is read-only during groupby

[erik-hasse]

• 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.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

import pandas as pd
df = pd.DataFrame(data={'x': [1], 'y': [2]})
df.to_parquet('pq_df', partition_cols='x')
df = pd.read_parquet('pq_df')
df.groupby('x', sort=False)

Problem description

The above code raises an exception:

Traceback (most recent call last):
  File "mve.py", line 5, in <module>
    df.groupby('x', sort=False)
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/frame.py", line 5798, in groupby
    return groupby_generic.DataFrameGroupBy(
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/groupby/groupby.py", line 402, in __init__
    grouper, exclusions, obj = get_grouper(
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/groupby/grouper.py", line 615, in get_grouper
    Grouping(
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/groupby/grouper.py", line 312, in __init__
    self.grouper, self.all_grouper = recode_for_groupby(
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/groupby/categorical.py", line 72, in recode_for_groupby
    cat = cat.add_categories(c.categories[~c.categories.isin(cat.categories)])
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/indexes/base.py", line 4667, in isin
    return algos.isin(self, values)
  File "/Users/ehasse/.local/lib/python3.8/site-packages/pandas/core/algorithms.py", line 447, in isin
    return f(comps, values)
  File "pandas/_libs/hashtable_func_helper.pxi", line 555, in pandas._libs.hashtable.ismember_int64
  File "stringsource", line 658, in View.MemoryView.memoryview_cwrapper
  File "stringsource", line 349, in View.MemoryView.memoryview.__cinit__
ValueError: buffer source array is read-only

This specifically requires the following:

• The dataframe is loaded from a parquet file.
• The column being grouped by was used to partition the file.
• sort=False is passed.

In addtion, passing observed=True stops the error from occurring.

I believe this is related to #31710, but they were unable to provide an example for groupby, and the issue remains on 1.0.3.

Expected Output

A DataFrameGroupBy object.

Output of 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 : 1.0.3
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 46.0.0.post20200309
Cython : None
pytest : 5.4.1
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.11.1
IPython : 7.13.0
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : 3.2.1
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 0.16.0
pytables : None
pytest : 5.4.1
pyxlsb : None
s3fs : None
scipy : 1.4.1
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None
numba : 0.47.0

[TomAugspurger]

Here's a reproducer without parquet.

In [29]: cats = np.array([1])

In [30]: cats.flags.writeable = False

In [31]: df = pd.DataFrame({"a": [1], "b": pd.Categorical([1], categories=pd.Index(cats))})

In [32]: df.groupby("b", sort=False)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-32-882468ddac04> in <module>
----> 1 df.groupby("b", sort=False)

~/sandbox/pandas/pandas/core/frame.py in groupby(self, by, axis, level, as_index, sort, group_keys, squeeze, observed)
   5825             group_keys=group_keys,
   5826             squeeze=squeeze,
-> 5827             observed=observed,
   5828         )
   5829

~/sandbox/pandas/pandas/core/groupby/groupby.py in __init__(self, obj, keys, axis, level, grouper, exclusions, selection, as_index, sort, group_keys, squeeze, observed, mutated)
    408                 sort=sort,
    409                 observed=observed,
--> 410                 mutated=self.mutated,
    411             )
    412

~/sandbox/pandas/pandas/core/groupby/grouper.py in get_grouper(obj, key, axis, level, sort, observed, mutated, validate)
    623                 in_axis=in_axis,
    624             )
--> 625             if not isinstance(gpr, Grouping)
    626             else gpr
    627         )

~/sandbox/pandas/pandas/core/groupby/grouper.py in __init__(self, index, grouper, obj, name, level, sort, observed, in_axis)
    310
    311                 self.grouper, self.all_grouper = recode_for_groupby(
--> 312                     self.grouper, self.sort, observed
    313                 )
    314                 categories = self.grouper.categories

~/sandbox/pandas/pandas/core/groupby/categorical.py in recode_for_groupby(c, sort, observed)
     69     # including those missing from the data (GH-13179), which .unique()
     70     # above dropped
---> 71     cat = cat.add_categories(c.categories[~c.categories.isin(cat.categories)])
     72
     73     return c.reorder_categories(cat.categories), None

~/sandbox/pandas/pandas/core/indexes/base.py in isin(self, values, level)
   4872         if level is not None:
   4873             self._validate_index_level(level)
-> 4874         return algos.isin(self, values)
   4875
   4876     def _get_string_slice(self, key: str_t, use_lhs: bool = True, use_rhs: bool = True):

~/sandbox/pandas/pandas/core/algorithms.py in isin(comps, values)
    452             comps = comps.astype(object)
    453
--> 454     return f(comps, values)
    455
    456

~/sandbox/pandas/pandas/_libs/hashtable_func_helper.pxi in pandas._libs.hashtable.ismember_int64()
    553 @cython.wraparound(False)
    554 @cython.boundscheck(False)
--> 555 def ismember_int64(int64_t[:] arr, int64_t[:] values):
    556     """
    557     Return boolean of values in arr on an

~/sandbox/pandas/pandas/_libs/hashtable.cpython-37m-darwin.so in View.MemoryView.memoryview_cwrapper()

~/sandbox/pandas/pandas/_libs/hashtable.cpython-37m-darwin.so in View.MemoryView.memoryview.__cinit__()

ValueError: buffer source array is read-only

There's a standard way to get cython to accept readonly arrays (adding const?) but I don't recall exactly.

[jbrockmendel]

in hashtable_func_helper.pxi.in L 209 {{c_type}}[:] arr needs to be changed to const {{c_type}}[:] arr

[TomAugspurger]

Thanks. @erik-hasse are you interested in making a PR with that change and tests?

[erik-hasse]

Sure. I haven't contributed to Pandas before, anything I should read before making the change and writing the test?

[TomAugspurger]

Our contributing doc page is apparent broken right, now doc/source/development/contributing.rst should have all the information you need.

[jorisvandenbossche]

Our contributing doc page is apparent broken right, now doc/source/development/contributing.rst should have all the information you need.

The dev docs are still (or again since a few days) working fine: https://pandas.pydata.org/docs/dev/development/contributing.html