Pycharm linter flags output of concat as "type", not DataFrame

[jebob]

Minimum display example

df1 = pd.DataFrame({'A': ['A0']}, index=[0])
df2 = pd.DataFrame({'A': ['A4']}, index=[4])
result = pd.concat([df1, df2])

print(df1.loc[4])  # Pycharm doesn't flag
print(result.loc[4])  # Pycharm flags as "Unresolved attribute reference '.loc' for class 'type'"
                      # But runs fine anyway

Problem description

Pycharm's linter believes pd.concat as returning class "type", when it normally returns a dataframe.

I suspect that this is due to core.reshape.concat._Concatenator.get_result() using core.dtypes.concat._get_frame_result_type which sometimes returns SparseDataFrame (a type) and sometimes an object (normally a Dataframe?).

Extract from pandas.core.dtypes.concat

def _get_frame_result_type(result, objs):
    """
    return appropriate class of DataFrame-like concat
    if any block is SparseBlock, return SparseDataFrame
    otherwise, return 1st obj
    """
    if any(b.is_sparse for b in result.blocks):
        from pandas.core.sparse.api import SparseDataFrame
        return SparseDataFrame
    else:
        return objs[0]

Expected Output

The MDE should not throw an error in Pycharm.

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.6.2.final.0 python-bits: 64 OS: Windows OS-release: 10 machine: AMD64 processor: Intel64 Family 6 Model 94 Stepping 3, GenuineIntel byteorder: little LC_ALL: None LANG: None LOCALE: None.None

pandas: 0.20.3
pytest: None
pip: 9.0.1
setuptools: 28.8.0
Cython: None
numpy: 1.13.1
scipy: None
xarray: None
IPython: None
sphinx: None
patsy: None
dateutil: 2.6.1
pytz: 2017.2
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: None
openpyxl: 2.4.8
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: 4.6.0
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: None
s3fs: None
pandas_gbq: None
pandas_datareader: None
None

I am using Pycharm Community Edition 2017.1.5.

[jreback]

you should address this to PyCharm.

[jebob]

Link to Pycharm issue: https://youtrack.jetbrains.com/issue/PY-25326

[gfyoung]

@jebob : The Pycharm linter is prone to false positives like this because this is not Java (their first language under the IntelliJ suite) 😄 . In fact, some of the warnings that I've seen similar to the one that you've reported are complete nonsense.

This is not to say that the linter doesn't get it right some times, but generally, flake8 is your friend here and so is running the code manually.

[jebob]

@gfyoung I've been suppressing the warnings in my code using type hints instead.
foo = pd.concat([bar1, bar2]) # type: pd.DataFrame

[gfyoung]

That works too, just hopefully you don't have to make too many hints 😄

[jebob]

Pycharm reads this piece of the documentation from DataFrame.concat().

   Returns
    -------
    concatenated : type of objects

and concludes that the return value is type type. This could be easily resolved on this end by using

   Returns
    -------
    concatenated : object, type of objs

which is technically correct, equally informative to the naked eye, solves the issue AND fixes a typo (objects vs objs).

[jreback]

a doc change like that would be fine