BUG: Pandas any() returning false with true values present (GH #23070)

[makbigc]

• closes Pandas any() returning false with true values present #23070
• 1 test added
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[jreback]

you can just make this:

data = self

no need for other things here

[jreback]

can you remove the commented code above (or see if it does anything)

[codecov]

Codecov Report

Merging #24434 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #24434      +/-   ##
==========================================
- Coverage    92.3%    92.3%   -0.01%     
==========================================
  Files         163      163              
  Lines       51987    51982       -5     
==========================================
- Hits        47989    47984       -5     
  Misses       3998     3998

Flag	Coverage Δ
#multiple	90.71% <100%> (-0.01%)	⬇️
#single	42.99% <33.33%> (-0.01%)	⬇️

Impacted Files	Coverage Δ
pandas/core/internals/managers.py	95.94% <100%> (+0.01%)	⬆️
pandas/core/frame.py	96.91% <100%> (ø)	⬆️
pandas/core/generic.py	96.63% <100%> (ø)	⬆️
pandas/core/tools/datetimes.py	85.58% <0%> (-0.18%)	⬇️
pandas/core/arrays/interval.py	93.02% <0%> (-0.03%)	⬇️
pandas/core/arrays/base.py	97.59% <0%> (-0.02%)	⬇️
pandas/core/indexes/datetimelike.py	97.58% <0%> (-0.01%)	⬇️
pandas/core/arrays/categorical.py	95.34% <0%> (-0.01%)	⬇️
pandas/core/indexes/multi.py	95.58% <0%> (-0.01%)	⬇️
pandas/core/indexes/base.py	96.27% <0%> (-0.01%)	⬇️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 08c920e...85a702d. Read the comment docs.

[codecov]

Codecov Report

Merging #24434 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #24434      +/-   ##
==========================================
- Coverage   92.31%   92.31%   -0.01%     
==========================================
  Files         166      166              
  Lines       52335    52412      +77     
==========================================
+ Hits        48313    48382      +69     
- Misses       4022     4030       +8

Flag	Coverage Δ
#multiple	90.73% <100%> (-0.01%)	⬇️
#single	43.05% <0%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/frame.py	96.91% <100%> (ø)	⬆️
pandas/core/indexes/datetimelike.py	96.23% <0%> (-1.37%)	⬇️
pandas/core/indexes/period.py	92.46% <0%> (-0.23%)	⬇️
pandas/core/arrays/datetimelike.py	95.5% <0%> (-0.16%)	⬇️
pandas/core/arrays/datetimes.py	97.74% <0%> (-0.03%)	⬇️
pandas/core/indexes/datetimes.py	96.33% <0%> (+0.19%)	⬆️
pandas/core/indexes/timedeltas.py	90.65% <0%> (+0.4%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update aeff38d...9007017. Read the comment docs.

[jreback]

in double-backticks say axis=1 and data type is datetimelike

[makbigc]

The test_any_all is failed in python2. An error is raised in line 239.

pandas/pandas/tests/frame/test_analytics.py

Lines 231 to 240 in 70bc919

	class NonzeroFail(object):
	def __nonzero__(self):
	raise ValueError
	mixed['_nonzero_fail_'] = NonzeroFail()
	if has_bool_only:
	getattr(mixed, opname)(axis=0, bool_only=True)
	getattr(mixed, opname)(axis=1, bool_only=True)

1. Reason why test_any_all is passed in python3, but failed in python2
   The class NonzeroFail implements the __nonzero__ method. In python2, bool() calls __nonzero__ method which raises the error. In python3, bool calls __bool__ and then __len__. If none of the above methods were implemented, True is returned. https://docs.python.org/3/reference/datamodel.html

2. Reason why test_any_all is failed in this PR
   self is given to data, instead of self._get_bool_data. _get_bool_data gets rid of the object block of NonzeroFail.

pandas/pandas/core/frame.py

Lines 7293 to 7294 in 70bc919

	elif filter_type == 'bool':
	data = self._get_bool_data()

The test with DataFrame containing NonzeroFail was passed due to the bug we are fixing. So in the comment block we deleted, the writer had been surprised by that the test with NonzeroFail would pass. The NonzeroFail is removed from the test.

[jreback]

thanks @makbigc