`concat` without `sort` in groupby.py code raises FutureWarning

[aloukina]

Problem description

There are several calls to concat in groupby.py that do not pass the sort argument. This raises FutureWarning when the function is exectuted. For example:

pandas/core/groupby/groupby.py:1105: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version 
of pandas will change to not sort by default.

To accept the future behavior, pass 'sort=False'.

To retain the current behavior and silence the warning, pass 'sort=True'.

  result = concat(values, axis=self.axis)

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.8.final.0
python-bits: 64
OS: Darwin
OS-release: 17.7.0
machine: x86_64
processor: i386
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.23.4
pytest: None
pip: 18.1
setuptools: 40.6.3
Cython: None
numpy: 1.14.2
scipy: 1.1.0
pyarrow: None
xarray: None
IPython: 6.5.0
sphinx: 1.8.2
patsy: 0.5.1
dateutil: 2.7.5
pytz: 2018.9
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.1.2
openpyxl: 2.5.12
xlrd: 1.2.0
xlwt: 1.2.0
xlsxwriter: None
lxml: None
bs4: 4.7.1
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[WillAyd]

Can you share the code you have to reproduce? I see you are using v0.23.4 this has most likely been fixed in more recent release

[aloukina]

Here is the corresponding line in master:

pandas/pandas/core/groupby/groupby.py

Line 887 in e3b0950

result = concat(values, axis=self.axis)

[blu3r4y]

There are still a lot of calls that are missing the sort argument in groupby.py on master, triggering internal warnings.

The last warning I got in 0.25.1 was here: (it's still the same line as mentioned by @aloukina)

pandas/pandas/core/groupby/groupby.py

Line 937 in 760c654

result = concat(values, axis=self.axis)

[TomAugspurger]

@blu3r4y @aloukina do either of you have a reproducible example that generates the error? That'll help in getting it addressed.

[blu3r4y]

@TomAugspurger sure! (tested with pandas==0.25.1)

import pandas as pd

df = pd.DataFrame({
    "group": ["a", "b", "c"],
    "value": [1, 2, 3]
})

df.groupby("group")["value"].apply(pd.DataFrame)

This yields the FutureWarning in groupby.py:937

pandas/pandas/core/groupby/groupby.py

Line 937 in 81fe56a

result = concat(values, axis=self.axis)

[TomAugspurger]

Thanks! That one is coming from https://github.com/pandas-dev/pandas/blob/c4489cbfa5f93e69301486e1ef867caed9f4eb9e/pandas/core/groupby/groupby.py#L939 . Anyone interested in making a PR?

…

On Mon, Sep 30, 2019 at 6:02 AM Mario Kahlhofer ***@***.***> wrote: @TomAugspurger <https://github.com/TomAugspurger> sure! (tested with pandas==0.25.1) import pandas as pd df = pd.DataFrame({ "group": ["a", "b", "c"], "value": [1, 2, 3] }) df.groupby("group")["value"].apply(pd.DataFrame) This yields the FutureWarning in groupby.py:937 https://github.com/pandas-dev/pandas/blob/81fe56a53a67095979764581aca0cab623b941eb/pandas/core/groupby/groupby.py#L937 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#25174?email_source=notifications&email_token=AAKAOIR7XHINRFRBIUVSII3QMHMFBA5CNFSM4GURF6EKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD75IRDA#issuecomment-536512652>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAKAOIT6RSZ2ESVCTE6TAKTQMHMFBANCNFSM4GURF6EA> .

[jbrockmendel]

Is this closed by #29786?

[jbrockmendel]

I don't see a warning in master, so closing as closed by #29786. Feel free to ping if the issue reappears and we can re-open.