handling of data frames with matrix columns
Closed this issue · 1 comments
adibender commented
For more complex data sets (with matrix columns as required to fit cumulative effects), dplyr verbs often don't work
Error: Each variable must be a 1d atomic vector or list
May be worth writing functions that clean the data set, i.e. transform 1d vectors to atomic (e.g. function clean.ped
) and/or call function after checking for and excluding these elements.
Another possibility would be to store matrix colums as list and only restore the matrix format before fitting the model and vice versa (as_clean_ped
, as_fit_ped
)
fabian-s commented
Using list columns seems to me to be the most hadlyeverse-compliant solution.
or maybe even use tidyr
's nested tables?