how to define a DataFrameModel that can have arbitrary number of columns but each of them should be float.

[rja172]

I have a function that can return arbitrary number of columns with the only constraint that each of them will be float. How can I create a DataFrameModel that enforces this and works nicely with mypy

class FloatSchema(pa.DataFrameModel):
   # i need to define something like this 
   # *: pa.Series[float]
   pass

def tranform(n_columns: int = 10) -> FloatSchema:
     data = ...
     df = pd.DataFrame(data)
     return df.pipe(FloatSchema)

[cosmicBboy]

Hi @rja172 you can use the regex=True combined with alias=... arguments

class FloatSchema(pa.DataFrameModel):
   col: pa.Series[float] = pa.Field(regex=True, alias=".+")

Or you can specify the type at the dataframe-level with a Config class

class FloatSchema(pa.DataFrameModel):
    class Config:
        dtype = float

[ragrawal]

oh this is awesome.. Thanks