BUG: GroupBy.apply with as_index=False still produces a MultiIndex · Issue #58291 · pandas-dev/pandas (original) (raw)

Pandas version checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of pandas.
• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

Assume we have this data:

In [88]: df = pd.DataFrame([ ...: ["group_a", 0], ...: ["group_a", 2], ...: ["group_b", 1], ...: ["group_b", 3], ...: ["group_b", 5], ...: ], columns=["group", "value"]) ...: ...: df In [88]: def up_to_two_rows(df: pd.DataFrame) -> pd.DataFrame: ...: return df.head(2)

Calling .apply seems to always create a MultiIndex, even when you don't ask for the groups:

In [89]: df.groupby("group").apply(up_to_two_rows)
<ipython-input-89-e4954502d06d>:1: DeprecationWarning: DataFrameGroupBy.apply operated on the grouping columns. This behavior is deprecated, and in a future version of pandas the grouping columns will be excluded from the operation. Either pass `include_groups=False` to exclude the groupings or explicitly select the grouping columns after groupby to silence this warning.
  df.groupby("group").apply(up_to_two_rows)
Out[89]: 
             group  value
group                    
group_a 0  group_a      0
        1  group_a      2
group_b 2  group_b      1
        3  group_b      3

In [92]: df.groupby("group").apply(up_to_two_rows, include_groups=False)
Out[92]: 
           value
group           
group_a 0      0
        1      2
group_b 2      1
        3      3

In [93]: df.groupby("group", as_index=False).apply(up_to_two_rows, include_groups=False)
Out[93]: 
     value
0 0      0
  1      2
1 2      1
  3      3

Issue Description

I am ultimately trying to get output that matches:

In [97]: df.groupby("group").apply(up_to_two_rows, include_groups=False).reset_index(level=0)
Out[97]: 
     group  value
0  group_a      0
1  group_a      2
2  group_b      1
3  group_b      3

But am not clear what combination of keywords is supposed to do that, if any

Expected Behavior

In [97]: df.groupby("group").apply(up_to_two_rows, include_groups=False).reset_index(level=0)
Out[97]: 
     group  value
0  group_a      0
1  group_a      2
2  group_b      1
3  group_b      3

Installed Versions

main