New way of working row-wise in R via purrr package

mutate_via_pmap_purrr.R

#' Change value of a column (\code{remarks}) for rows with some fields equal to some rows in a second df (called here \code{df_ref}).
#' Notice that all \code{pmap()} family of functions demands same number of columns as the input df. 
#' It is strongly recommended to use the column names as inputs of the function of \code{pmap()}.
#' In this specific example we require that \code{col3} is of type \code{character} and we use \code{pmap_chr()} to return a character.

library(purrr)
library(dplyr)

df <- 
   df %>%
   mutate(col3 = pmap_chr(df %>%
                               select(col1, col2, col3), 
     function(col1, col2, col3) {
       ifelse(all(c(col1, col2) %in% df_ref),
              paste(na.omit(col3), "Add this"),
              na.omit(col3))
       })
   )