Currently, transforms that rely on var_names_g (ZScore and DivideByScale) throw an error if input var_names_g doesn't match self.var_names_g. For example, if the Filter transform is used upstream then these transforms need to be instantiated with filtered __init__ arguments which makes it inconvenient to use. This issue proposes to make this transforms flexible and adapt to filtered or re-ordered inputs.
Currently, transforms that rely on
var_names_g
(ZScore
andDivideByScale
) throw an error if inputvar_names_g
doesn't matchself.var_names_g
. For example, if theFilter
transform is used upstream then these transforms need to be instantiated with filtered__init__
arguments which makes it inconvenient to use. This issue proposes to make this transforms flexible and adapt to filtered or re-ordered inputs.