Right now it is only going to work on tsibbles with a single measured variable. I would like to be able to pass it a variable name and have it create columns with names in the format of var_name.component_name. This way, we could viably pass it a series of variables and have the decomposition components for each appended.
I'm not sure if the best answer here would be to push it long based on the initial measured_vars and then bind_cols for the components, or just let the thing go super-wide. I would think the initial might just be superwide, arranging the columns so that the measured_var precedes the components.
Right now it is only going to work on tsibbles with a single measured variable. I would like to be able to pass it a variable name and have it create columns with names in the format of
var_name.component_name. This way, we could viably pass it a series of variables and have the decomposition components for each appended.I'm not sure if the best answer here would be to push it long based on the initial
measured_varsand thenbind_colsfor the components, or just let the thing go super-wide. I would think the initial might just be superwide, arranging the columns so that themeasured_varprecedes the components.