Loop across rows and colums with nested data - r

Question

I have the following data structure: Meetings in Persons in Groups. The groups met differently often and the number of group members varied for every meeting.
$ GroupID : chr "1" "1" "1" "1" ...
$ groupnames : chr "A&M" "A&M" "A&M" "A&M" ...
$ MeetiID : chr "1" "1" "2" "2" ...
$ Date_Meetings : chr "43293" "43293" "43298" "43298" ...
$ PersonID : num 171 185 171 185 185 113 135 113 135 113 ...
$ v_165 : chr "3" "3" "4" "3" ...
$ v_166 : chr "2" "2" "3" "3" ...
$ v_167 : chr "2" "4" "4" "3" ...
$ v_168 : chr "6" "7" "4" "5" ...
$ problemtypes_categories: chr "Knowledgeproblem" "Knowledgeproblem" "Motivationalproblem" "Coordinationproblem" ...
$ v_165_dicho : num 0 0 0 0 1 1 1 0 0 1 ...
$ v_166_dicho : num 0 0 0 0 0 0 0 0 0 0 ...
$ v_167_dicho : num 0 0 0 0 1 1 0 0 0 0 ...
Now I have to create a new variable that should be binary (0/1) with the name agreement_levels. So, every time, a person in one group has - regarding the same learning meeting - a same problem type category than the other learner(s) of the same group at the same meeting, both learners (or three or four, depending on the group size for a respective meeting) should get the value 1 at the agreement variable, else they should all get 0. Whenever a person (e.g., among four learners) already has a different category of problem than the others, there is a 0 on the agreement variable for all.
If only 1 person is in the data set for one and the same meeting, there must be a NA at agree. When one person has NA at the problemtype variable, however, and there are 2 people in the data set for the same meeting, both get 0 at agree; but if there are 4 people for the same meeting in the data set and one of them has NA at problemtype, then only this person but not the others get NA at agree.
I did already write a command, but it is not working yet and still does not consider the NAs:
GroupID1 <- df$GroupID[1:nrow,]
TreffID1 <- df$TreffID[1:nrow,]
for(i in 1:(GroupID1 -1){
for(j in 1:(TreffID1 -1){
if(df[i, 3] == df[i+1, 3]-1){
if(df[i, 15] == df[i+1, 15]-1){
df[c(i, i+1), 28] <- 1,
df[c(i, i+1), 28] <- 0
Many thanks in advance.
dput(head(df))
structure(list(GroupID = c("1", "1", "1", "1", "1", "2"), TreffID = c("1", "1",
"2", "2", "3", "1"), PersonID = c(171, 185, 171, 185,
185, 113), problemtypen_oberkategorien = c("Verständnisprobleme",
"Verständnisprobleme", "Motivationsprobleme", "Motivationsprobleme",
"Motivationsprobleme", "Motivationsprobleme"), passung.exkl = c("0",
"0", "0", "0", "1", "1")), row.names = c(NA, -6L), class = c("tbl_df",
"tbl", "data.frame"))

Answer 1

Instead of loops, I used R's dplyr. I'm not sure if I got all your logic correct, since there was a lot there. For example, you didn't specify what would happen for NA problemtype and 3 people. But here is a starting point that uses group_by, so you are looking within each set of rows with the same GroupID and TreffID, and then mutate and case_when, which assign values to a new column, according to criteria, and then functions like n() that count how many rows and n_distinct that count distinct rows so you if it is ==1 then we know they are all the same.
library(tidyverse)
df <- df %>%
group_by(GroupID, TreffID) %>%
mutate(agreement_levels = case_when(n() == 1 ~ -1,
is.na(problemtypen_oberkategorien) & n() == 2 ~ 0,
is.na(problemtypen_oberkategorien) & n() > 2 ~ -1,
n_distinct(problemtypen_oberkategorien, na.rm = FALSE) == 1 ~ 1,
n_distinct(problemtypen_oberkategorien, na.rm = FALSE) > 1 ~ 0,
TRUE ~ -1),
agreement_levels = na_if(agreement_levels, -1)) %>%
select(GroupID, TreffID, problemtypen_oberkategorien, agreement_levels, everything())