TST: groupby of idxmax/min with mixed types

[Leejung8763]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

df.groupby('teamId').idxmax()
tmp.groupby('teamId').idxmax()

Problem description

DataFrame df is consist of teamId(int64) and neutralMinionsKilled(int64) and 'tmp' is consist of teamId(Int64) and neutralMinionsKilled(Int64)
df.groupby('teamId').idxmax() is working
But tmp.groupby('teamId').idxmax() is not working and showing 'TypeError: argmax() takes 1 positional argument but 2 were given'

Expected Output

Output of pd.show_versions()

[paste the output of pd.show_versions() here leaving a blank line after the details tag]

[erfannariman]

Can you add a reproducible example as mentioned when you opened the bug report, see link

[Leejung8763]

tmp = pd.DataFrame({'teamId': [100,100,100,200,200,200],
                   'value': [0,0,0,1,2,0],
                   })
tmp.groupby('teamId').idxmax()

It's working and the results are as follows.

But,

tmp = pd.DataFrame({'teamId': [100,100,100,200,200,200],
                   'value': [0,0,0,1,2,0],
                   })
tmp = tmp.convert_dtypes()
tmp.groupby('teamId').idxmax()

It's not working with "TypeError: argmax() takes 1 positional argument but 2 were given" message

[MarcoGorelli]

Thanks @Leejung8763 for the report

This works on master, but indeed fails in 1.1.3, will do a bisect

[MarcoGorelli]

Fixed in #38195

4749fd611e0a303b7325686901c90e6a362bf3cd is the first bad commit
commit 4749fd611e0a303b7325686901c90e6a362bf3cd
Author: jbrockmendel
Date:   Wed Dec 2 04:11:05 2020 -0800

    BUG: DataFrame.idxmin/idxmax with mixed dtypes (#38195)

Could do with an extra test for this though, pull requests welcome

[bsun94]

hi Marco, I'd love to take this on. I'm still somewhat new to pandas - hope you don't mind me reverting back if I have any questions!

[bsun94]

Actually, I think more tests have been added since the PR quoted above - looking at the test_reductions.py module, it seems more has been appended to test_idxmax_mixed_dtype since what's displayed in the file changes in the PR.