Add argument to `Series.value_counts` to set the name of the new column created about polars

[stinodego]

Oh yeah I think you're right..

[DeflateAwning]

Awesome, thanks! Good lesson about aliasing a Series's name though; didn't know I could do that :)

[cmdlineluser]

Yeah, it seems currently there is a hardcoded check for this potential situation:

df = pl.DataFrame({"count": ["a", "a", "b"]})

df["count"].value_counts()
# DuplicateError: using `value_counts` on a column named 'count' would lead to duplicate column names

It was recently added for GroupBy.len due to similar issues #15222

df["count"].to_frame().group_by("*").len("occurences")
# shape: (2, 2)
# ┌───────┬────────────┐
# │ count ┆ occurences │
# │ ---   ┆ ---        │
# │ str   ┆ u32        │
# ╞═══════╪════════════╡
# │ a     ┆ 2          │
# │ b     ┆ 1          │
# └───────┴────────────┘

[stinodego]

We do not add name parameters to operations on a single column.

DataFrame.with_row_index needs a name parameter because the name may clash with an existing column in the dataframe.

Series/Expr.value_counts don't need a name parameter because you can just use alias before:

df["count"].alias("my_custom_name").value_counts()

I'll add a hint to the error message.

[alexander-beedie]

Series/Expr.value_counts don't need a name parameter because you can just use alias before:

Actually, I think we probably should add a name param here, as we force the result name to "count". The user shouldn't have to change their column names to account for us forcing one on them 🤔 I'll make a PR with an example and you can see what you think.