define new column from list of columns in data.table rowwise

dt_new_col.R

# dt is a data.table with id_cols and value_cols
# we want rows where any of the value cols is not NA regardless of the 
# values of the id_cols

dt[, to_keep := purrr::pmap_lgl(
    purrr::map_df(
        dt[, .SD, .SDcols = value_cols],
        ~ !is.na(.x)
    ),
    any
)]