BUG: Timestamp array comparison may raise RecursionError

[sinhrks]

Code Sample, a copy-pastable example if possible

import numpy as np
import pandas as pd

n1 = np.array([pd.Timestamp('2011-01-01 00:00:00'),
               pd.Timestamp('2011-01-03 00:00:00-0500', tz='US/Eastern')], dtype=object)

pd.Timestamp('2011-01-03 00:00:00-0500', tz='US/Eastern') == n1 
# RecursionError: maximum recursion depth exceeded

Problem description

Related to #12780. This issue causes datetime-like .replace to raise SystemError. Recursion is caused by this line:

https://github.com/pandas-dev/pandas/blob/master/pandas/tslib.pyx#L1040

Expected Output

raises TypeError.

Output of pd.show_versions()

# Paste the output here pd.show_versions() here

[jreback]

that's not too friendly!

[jbrockmendel]

I think the error I just hit is related. In py 3.5.2 (pandas 0.20.2):

>>> periods = pd.Series([pd.Period('2018Q2', 'Q-DEC'), pd.Period('2021', '5A-DEC')], name='Period')
>>> df = pd.DataFrame(periods)
>>> df.set_index('Period', append=True)
pandas._libs.period.IncompatibleFrequency: Input has different freq=Q-DEC from Period(freq=5A-DEC)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.5/site-packages/pandas/core/frame.py", line 2836, in set_index
    index = MultiIndex.from_arrays(arrays, names=names)
  File "/usr/local/lib/python3.5/site-packages/pandas/core/indexes/multi.py", line 1100, in from_arrays
    labels, levels = _factorize_from_iterables(arrays)
  File "/usr/local/lib/python3.5/site-packages/pandas/core/categorical.py", line 2193, in _factorize_from_iterables
    return map(list, lzip(*[_factorize_from_iterable(it) for it in iterables]))
  File "/usr/local/lib/python3.5/site-packages/pandas/core/categorical.py", line 2193, in <listcomp>
    return map(list, lzip(*[_factorize_from_iterable(it) for it in iterables]))
  File "/usr/local/lib/python3.5/site-packages/pandas/core/categorical.py", line 2165, in _factorize_from_iterable
    cat = Categorical(values, ordered=True)
  File "/usr/local/lib/python3.5/site-packages/pandas/core/categorical.py", line 298, in __init__
    codes, categories = factorize(values, sort=True)
  File "/usr/local/lib/python3.5/site-packages/pandas/core/algorithms.py", line 567, in factorize
    assume_unique=True)
  File "/usr/local/lib/python3.5/site-packages/pandas/core/algorithms.py", line 486, in safe_sort
    sorter = values.argsort()
  File "pandas/_libs/period.pyx", line 723, in pandas._libs.period._Period.__richcmp__ (pandas/_libs/period.c:11713)
  File "/usr/local/lib/python3.5/site-packages/pandas/tseries/offsets.py", line 375, in __ne__
    return not self == other
  File "/usr/local/lib/python3.5/site-packages/pandas/tseries/offsets.py", line 364, in __eq__
    if isinstance(other, compat.string_types):
SystemError: <built-in function isinstance> returned a result with an error set

In the real-life version when the array of pd.Period objects is longer, the error message includes a ton of these:

During handling of the above exception, another exception occurred:

SystemError: <built-in function isinstance> returned a result with an error set

Halfway through the traceback is a call inside Categorical.__init__ to factorize(values, sort=True). That call is in a try/except that catches a TypeError and falls back to factorize(values, sort=False). Trying that out directly, factorize with sort=False appears just fine, so it is the sort=True that is causing the problem.

I don't do much with py3, so don't know what to make of the returned a result with an error set message. In py2 we get a NotImplementedError from the original set_index call. If we focus just on the factorize(values, sort=True) call, we get:

>>> values = periods
>>> codes, categories = factorize(values, sort=True)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python2.7/site-packages/pandas/core/algorithms.py", line 567, in factorize
    assume_unique=True)
  File "/usr/local/lib/python2.7/site-packages/pandas/core/algorithms.py", line 486, in safe_sort
    sorter = values.argsort()
  File "pandas/_libs/period.pyx", line 725, in pandas._libs.period._Period.__richcmp__ (pandas/_libs/period.c:11842)
pandas._libs.period.IncompatibleFrequency: Input has different freq=Q-DEC from Period(freq=5A-DEC)

I'll follow up on this later. For the purposes of this and related IncompatibleFrequency errors, I think defining an ordering has been avoided because there is not One Obvious Way. But for the purposes of making this Just Work, essentially any ordering would work.