Coercing a column of lists into a string in an R data frame

Question

Create sample data:

id <- c(12, 32, 42, 42, 52, 52, 67, 67)
relationship_id <- c(15,1,59,1,61,6,59,1)
sample.data <- data.frame(id,relationship_id)

For each id that appears more than once, concatenate the relationship_id:

combo <- aggregate(relationship_id ~ id, data = sample.data, paste, sep=",")
table(combo$relationship_id)
Error in table(combo$relationship_id) :
  all arguments must have the same length

I figured out what caused this error:

class(combo$relationship_id)
[1] "list"

But when I try and coerce the list vector to a character vector:

combo["relationship_id"] <- lapply(combo["relationship_id"], as.character)
> head(combo)    
  id relationship_id
1 12              15
2 32               1
3 42    c("59", "1")
4 52    c("61", "6")
5 67    c("59", "1")

It includes the concatenation syntax... I understand that I can parse the output so that it is usable, but why is this happening? Is there an easier way to clean up the output?

Answer 1

You are trying to tackle the wrong problem. If you really wanted to collapse those values into a single character vector, you should use collapse = "," instead of sep.

combo <- aggregate(relationship_id ~ id, data = sample.data, 
                   paste, collapse=",")
table(combo$relationship_id)
# 
#    1   15 59,1 61,6 
#    1    1    2    1