BUG: 1.3.0rc1 pd.util.hash_array on Index fails to access _values_for_factorize

[TheNeuralBit]

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

• I have confirmed this bug exists on the latest version of pandas. (does not exist in 1.2.4, only seen in 1.3.0rc1 and master)

• (optional) I have confirmed this bug exists on the master branch of pandas. (confirmed on 0b68d87)

---

Code Sample, a copy-pastable example

In [1]: import pandas as pd

In [2]: pd.util.hash_array(pd.DatetimeIndex(['2018-10-28 01:20:00'], tz='Europe/Berlin'))
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-2-389898c5af02> in <module>
----> 1 pd.util.hash_array(pd.DatetimeIndex(['2018-10-28 01:20:00'], tz='Europe/Berlin'))

~/working_dir/pandas/pandas/core/util/hashing.py in hash_array(vals, encoding, hash_key, categorize)
    287     elif not isinstance(vals, np.ndarray):
    288         # i.e. ExtensionArray
--> 289         vals, _ = vals._values_for_factorize()
    290 
    291     return _hash_ndarray(vals, encoding, hash_key, categorize)

AttributeError: 'DatetimeIndex' object has no attribute '_values_for_factorize'

In [3]: pd.util.hash_array(pd.Index([1,2,3]))
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-3-e66efa244441> in <module>
----> 1 pd.util.hash_array(pd.Index([1,2,3]))

~/working_dir/pandas/pandas/core/util/hashing.py in hash_array(vals, encoding, hash_key, categorize)
    287     elif not isinstance(vals, np.ndarray):
    288         # i.e. ExtensionArray
--> 289         vals, _ = vals._values_for_factorize()
    290 
    291     return _hash_ndarray(vals, encoding, hash_key, categorize)

AttributeError: 'Int64Index' object has no attribute '_values_for_factorize'

In [4]: pd.util.hash_array(pd.RangeIndex(1,3))
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-4-af0900aae979> in <module>
----> 1 pd.util.hash_array(pd.RangeIndex(1,3))

~/working_dir/pandas/pandas/core/util/hashing.py in hash_array(vals, encoding, hash_key, categorize)
    287     elif not isinstance(vals, np.ndarray):
    288         # i.e. ExtensionArray
--> 289         vals, _ = vals._values_for_factorize()
    290 
    291     return _hash_ndarray(vals, encoding, hash_key, categorize)

AttributeError: 'RangeIndex' object has no attribute '_values_for_factorize'

Problem description

This issue looks similar to #41817, but that is specifically for DateTimeIndex with tz defined, while this seems to happen for any Index instance.

Expected Output

A hash of the input index, as in pandas < 1.3.0.

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit : 2dd9e9b python : 3.8.6.final.0 python-bits : 64 OS : Linux OS-release : 5.10.28-1rodete1-amd64 Version : #1 SMP Debian 5.10.28-1rodete1 (2021-04-30) machine : x86_64 processor : byteorder : little LC_ALL : None LANG : en_US.UTF-8 LOCALE : en_US.UTF-8

pandas : 1.3.0rc1
numpy : 1.19.5
pytz : 2021.1
dateutil : 2.8.1
pip : 20.2.1
setuptools : 49.2.1
Cython : 0.29.22
pytest : 6.2.2
hypothesis : 6.4.0
sphinx : 3.5.1
blosc : 1.10.2
feather : None
xlsxwriter : 1.3.7
lxml.etree : 4.6.2
html5lib : 1.1
pymysql : None
psycopg2 : None
jinja2 : 2.11.3
IPython : 7.21.0
pandas_datareader: None
bs4 : 4.9.3
bottleneck : 1.3.2
fsspec : 0.8.7
fastparquet : 0.5.0
gcsfs : 0.7.2
matplotlib : 3.3.4
numexpr : 2.7.3
odfpy : None
openpyxl : 3.0.6
pandas_gbq : None
pyarrow : 3.0.0
pyxlsb : None
s3fs : 0.5.2
scipy : 1.6.1
sqlalchemy : 1.3.23
tables : 3.6.1
tabulate : 0.8.9
xarray : 0.17.0
xlrd : 2.0.1
xlwt : 1.3.0
numba : 0.52.0

[jbrockmendel]

hash_array is only supposed to get ndarray or ExtensionArray (see the annotation)

[TheNeuralBit]

(see the annotation)

I'm sorry I'm not sure what annotation you're referring to?

It does look this is WAI though, and I should prefer pd.util.hash_pandas_object for my application.

Would a cleaner error message be a welcome change here?

[jbrockmendel]

I'm sorry I'm not sure what annotation you're referring to?

The annotation for values on hash_array is ArrayLike, which is an alias for np.ndarray | ExtensionArray.

and I should prefer pd.util.hash_pandas_object for my application

Yes, that should accept Indexes just fine.

[jbrockmendel]

Would a cleaner error message be a welcome change here?

sure

[simonjayhawkins]

This issue looks similar to #41817, but that is specifically for DateTimeIndex with tz defined, while this seems to happen for any Index instance.

the code sample in the OP includes a tz. so that is not a change in 1.3.0rc1

[simonjayhawkins]

The public api for hash_array is at https://pandas.pydata.org/docs/reference/api/pandas.util.hash_array.html

The docstring has been updated in #39949 and the api widened to accept any EA and numpy array. https://pandas.pydata.org/pandas-docs/version/1.3/reference/api/pandas.util.hash_array.html

so Index is not an accepted input. but maybe we should also add something to the 1.3 release notes since it worked for some Index types before. @jbrockmendel?

[simonjayhawkins]

• does not exist in 1.2.4, only seen in 1.3.0rc1 and master

first bad commit: [9664284] TYP: fix ignores (#40452)

[jbrockmendel]

maybe we should also add something to the 1.3 release notes since it worked for some Index types before

makes sense