pd.merge() on int+float produces object since 0.20

[sobayed]

Code Sample, a copy-pastable example if possible

In [2]: df1 = pd.DataFrame({'key': [1, 2], 'value': [0, 1]})

In [3]: df1.dtypes
Out[3]: 
key      int64
value    int64
dtype: object

In [4]: df2 = pd.DataFrame({'key': [1.0, 2.0], 'other_value': ['A', 'B']})

In [5]: df2.dtypes
Out[5]: 
key            float64
other_value     object
dtype: object

In [6]: print(pd.merge(df1, df2, how='left', on='key').dtypes)
key            object
value           int64
other_value    object
dtype: object

Problem description

I was expecting that in the merged DataFrame's "key" column pandas would either upcast int to float (like it does e.g. when missing values occur in an int column) or leave the column dtype as int.

I checked and confirmed that the latter was the behaviour in pandas 0.19.2

Expected Output

Merging two DataFrames where the key column is of type int in the one and of type float in the other frame produces a key column of type int or float in the resulting frame (not object).

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.3.final.0
python-bits: 64
OS: Linux
OS-release: 4.4.0-81-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.21.0
pytest: 3.2.1
pip: 9.0.1
setuptools: 36.5.0.post20170921
Cython: 0.26.1
numpy: 1.13.3
scipy: 1.0.0
pyarrow: None
xarray: None
IPython: 6.1.0
sphinx: 1.6.3
patsy: 0.4.1
dateutil: 2.6.1
pytz: 2017.2
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.2
feather: None
matplotlib: 2.1.0
openpyxl: 2.4.8
xlrd: 1.1.0
xlwt: 1.3.0
xlsxwriter: 1.0.2
lxml: 4.1.0
bs4: 4.6.0
html5lib: 0.999999999
sqlalchemy: 1.1.13
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[mroeschke]

Looks fixed on master. Could use a test.

In [25]: print(pd.merge(df1, df2, how='left', on='key').dtypes)
key             int64
value           int64
other_value    object
dtype: object

In [26]: pd.__version__
Out[26]: '0.25.0.dev0+145.g9c0f6a8d7'

[rbenes]

Duplicity of this: #16572?
Fixed in #18352 even with tests...

[WillAyd]

Thanks for finding that @rbenes