[BUG] df.groupby(['a']).agg({'b': ['mean', 'mean']}) only produces a single column (as opposed to 2)

[MarcoGorelli]

Describe the bug
A clear and concise description of what the bug is.

Steps/Code to reproduce bug

df = cudf.DataFrame({'a': [1,1,2], 'b': [4,5,6]})
df.groupby(['a']).agg({'b': ['mean', 'mean']})

outputs

 	b
	mean
a 	
1 	4.5
2 	6.0

Expected behavior

what pandas does:

 	b
	mean 	mean
a 		
1 	4.5 	4.5
2 	6.0 	6.0

Spotted in Narwhals

Environment overview (please complete the following information)

• Environment location: [Bare-metal, Docker, Cloud(specify cloud provider)]
• Method of cuDF install: [conda, Docker, or from source]
  • If method of install is [Docker], provide docker pull & docker run commands used

24.10.01, google colab

Environment details
Please run and paste the output of the cudf/print_env.sh script here, to gather any other relevant environment details

Additional context
Add any other context about the problem here.

[galipremsagar]

To support this case we will need to add support for duplicate column names in cudf, which I don't will be supported in near future. @mroeschke does pandas intend to continue supporting duplicate column name use-case?

[MarcoGorelli]

thanks for your response

tbf now that you've added support for NamedAgg in cuDF, we can probably rewrite how we do group_by in Narwhals for pandas-like libraries and avoid the duplicate column names completely

@mroeschke does pandas intend to continue supporting duplicate column name use-case?

the last I remember hearing about this was that duplicate column names are a fact of life and should continue to be supported (though to be honest I think they create more problems than they solve)

[mroeschke]

Yeah agreed with Marco that I think pandas will continue to support duplicate column names in the foreseeable future

[vyasr]

Yes that is consistent with what I've heard in previous discussions about this as well.

[MarcoGorelli]

Same for cudf.pandas for this issue

OK with just raising here on cudf's side?