Accessing a columns name via `.alias` makes the type checker complain

[FrithjofWinkelmannQC]

When accessing the name of a column via .alias, the type checker complains, since the type is str | None. See the following example:

class MetricSchema(dy.Schema):
    name = dy.String()
    value = dy.Float64()

df.join(other_df, on=MetricSchema.name.alias)
# mypy will complain with something like
# error: Argument 2 to "join" of "Expr" has incompatible type "str | None"; expected "str"

Right now you either have to silence mypy with eg # type: ignore[arg-type] or define an extra string constant for column names. It would be nice to have a method similar to .col() that raises a ValueError if the column name is not set and otherwise returns the string name.

[borchero]

@FrithjofWinkelmannQC: #52 introduces a .name parameter that you can use. The use of .alias is discouraged and it won't be filled if you don't specify the alias= parameter when initializing columns.

We will cut a new release at the end of the week :)

[FrithjofWinkelmannQC]

Awesome, thanks for adressing this :)