BUG/CLN?: get_duplicates in _build_names_mapper

[twoertwein]

This function looks like it doesn't do what it is intended to do:

pandas/pandas/core/reshape/pivot.py

Line 864 in 5c4cc4a

def get_duplicates(names):

Based on the implementation it can be replaced by just set(names) but that doesn't match with the intention of the name.

[samjeffrees]

Hi @twoertwein,

I've just had a look and based on the docstring in _build_names_mapper() and it looks like this isn't working as intended. This can be fixed by changing get_duplicates to something like the following:

def get_duplicates(names):
    seen: set = set()
    duplicate: set = set()
    for name in names:
        if name not in seen:
            seen.add(name)
        else:
            duplicate.add(name)
    return duplicate

Is it okay if I pick this up?

I'm new here so pls forgive the ignorance :-)

EDIT: I'm going to take this, better to seek forgiveness etc.

[samjeffrees]

take

[twoertwein]

This can be fixed by changing get_duplicates to something like the following

Sorry, I can't help you with that - I'm not familair with that code, just stumbled into it accidentaly.

[topper-123]

This does look weird. @samjeffrees, could you try your solution and and see if it breaks some tests?

[tqa236]

take