How to implement type hinting for DataFrame

Question

I would like to have same sort of Python type hinting on Polars DataFrame. Since DataFrame is not a generic type, I cannot do DataFrame[MyFrameSchema].

For example:

MyFrameSchema = {'price': float, 'quantity': int, 'product_name': str}

or:

class MyDataFrame(pl.DataFrame)
    ...

I now Polars has type SchemaDict in DataFrame constructor from polars.type_aliases import SchemaDict, but I don't know how to use it in Python code.

Maybe someone has idea how to wrap DataFrame into TMyDataFrame Something like TMyDataFrame: TypeAlias = ... or something else. I am open for suggestions.

Answer 1

There are two Polars DataFrame validators that I know of, pandera (https://github.com/unionai-oss/pandera) and patito (https://github.com/JakobGM/patito). However, these do runtime validation and aren't really type hints.