Validate multiple outputs using 'check_types'

[fimselamse]

Currently, I am only able to validate multiple outputs using check_io:

@pa.check_io(out=[(0, Schema0), (1, Schema1), (2, Schema2)])
def func(a,b,c): -> Tuple[DataFrame,DataFrame,DataFrame]
    ....
    return df0, df1, df2

However, this seems like a misuse of check_io, as I'm not checking for input. Instead, I would expect check_types to handle multiple outputs:

@pa.check_types(lazy=True)
def func(a,b,c): -> Tuple[DataFrame[Schema0],DataFrame[Schema0],DataFrame[Schema0]]
    ....
    return df0, df1, df2

Additionally, it seems check_io only lazy-checks one output at a time before failing, instead of checking them all before returning the errors.

The obvious workaround is to split the function into three, although this would create some redundancy if they share logic.