Objects' IDs by common value in different variables - r

Question

I have a dataset with self-ratings and peer-ratings. The dataset is in long format. Before reshaping the dataset into wide-format, I want to give self-ratings and peer-ratings a common ID so that I can later match the peer-ratings to the self-ratings by that ID. The data look like this:
| questionnaire | ID | REF | SERIAL | x | y |
|---------------|----|------|--------|----|----|
| self | 1 | 1234 | NA | 4 | NA |
| self | 2 | 2345 | NA | 6 | NA |
| peer | NA | NA | 1234 | NA | 8 |
| peer | NA | NA | 2345 | NA | 4 |
The self-ratings have a reference variable ("REF") which refer to a peer-rating. The peer-ratings have the same value in the variable "SERIAL".
I'm now trying to attribute the same ID to the peer-ratings as the ID of the self-ratings which refer to the peers by the SERIAL value. The table should look like this then:
| questionnaire | ID | REF | SERIAL | x | y |
|---------------|----|------|--------|----|----|
| self | 1 | 1234 | NA | 4 | NA |
| self | 2 | 2345 | NA | 6 | NA |
| peer | 1 | NA | 1234 | NA | 8 |
| peer | 2 | NA | 2345 | NA | 4 |
How could I do this best?

Answer 1

may I can help you. I would like to change the ID values right in the data frame. For that I would do a loop around the rows by picking up an ID from the right counterpart and adding it into the NA's in the data frame. In my following code example, I've searched for the right ID for every peer. The map_df function does the loop and prepares as a output a data frame.
library(purrr)
library(dplyr)
marcel.data <- data.frame(questionnaire = c("self", "self", "peer", "peer"),
ID = c(1, 2, NA, NA),
REF = c("1234", "2345", NA, NA),
SERIAL = c(NA, NA, "1234", "2345"),
x = c(4, 6, NA, NA),
y = c(NA, NA, 8, 4))
new_id_data <- map_df(1:nrow(marcel.data), function(i){
if(marcel.data[i, 1] == "peer") { # matching peer to Self
serial <- unique(marcel.data[i, "SERIAL"])
new.id <- marcel.data %>% # getting the needed ID
filter(REF == serial & questionnaire == "self") %>%
select(ID)
marcel.data[i, "ID"] <- new.id # adding the new ID to the old data
marcel.data[i,]
} else
(marcel.data[i,])
})