Fix Issue #34923: Inferred dtype at the end of df explode method

[SanthoshBala18]

• closes Infer dtype when using df.explode()ENH: #34923
• tests added / passed
  pandas/tests/frame/methods/test_explode.py, method test_inferred_dtype
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[pep8speaks]

Hello @SanthoshBala18! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2020-07-20 11:26:43 UTC

[jreback]

you may need to update the example in the docs / doc-string as well

[jreback]

lgtm pls merge master and ping on green

[SanthoshBala18]

@jreback, merged the master branch..

[simonjayhawkins]

I'm -1 on this. This is a breaking change and not a bug.

the change for lists of ints with some rows containing empty lists now being cast to float should wait till a major release and maybe use the nullable integer type instead.

current

>>> pd.__version__
'1.1.0rc0'
>>>
>>> s = pd.Series([[1, 2, 3, 4], [], [5, 6, 7]])
>>> s
0    [1, 2, 3, 4]
1              []
2       [5, 6, 7]
dtype: object
>>>
>>> res = s.explode()
>>> res
0      1
0      2
0      3
0      4
1    NaN
2      5
2      6
2      7
dtype: object
>>>
>>> type(res.iloc[1])
<class 'int'>
>>>

this PR

>>> pd.__version__
'1.1.0rc0+13.gf6144512d'
>>>
>>> s = pd.Series([[1, 2, 3, 4], [], [5, 6, 7]])
>>> s
0    [1, 2, 3, 4]
1              []
2       [5, 6, 7]
dtype: object
>>>
>>> res = s.explode()
>>> res
0    1.0
0    2.0
0    3.0
0    4.0
1    NaN
2    5.0
2    6.0
2    7.0
dtype: float64
>>>
>>> type(res.iloc[1])
<class 'numpy.float64'>
>>>

[jreback]

closing this. see my comments on the issue (basically need to add an option downcast= instead here.