Count unique values across columns in R

Question

I am trying to create a new variable with unique counts of string values from two different columns. So I have something like this, for example:

# A tibble: 4 x         


        

Answer 1

Here is a method using tidyverse without looping

library(tidyverse)
df1 %>% 
   mutate(partners = str_c(names, partners, sep=", ")) %>%
   separate_rows(partners) %>%
   distinct %>% 
   count(names) %>% 
   right_join(df1)
# A tibble: 4 x 3
#  names       n partners                 
#  <fct>   <int> <fct>                    
#1 John        4 Mary, Ashley, John, Kate 
#2 Mary        3 Charlie, John, Mary, John
#3 Charlie     3 Kate, Marcy              
#4 David       3 Mary, Claire         

Answer 2

There's another way with toString.

dat$uniquecounts <- sapply(strsplit(apply(dat, 1, toString), ", "), 
                           function(x) length(unique(x)))

dat
#     names                  partners uniquecounts
# 1    John  Mary, Ashley, John, Kate            4
# 2    Mary Charlie, John, Mary, John            3
# 3 Charlie               Kate, Marcy            3
# 4   David              Mary, Claire            3

Answer 3

Withtidyverse, first convert factor columns to character, use map2 and split partners to individual vector of strings and then count unique values combining with names using n_distinct.

library(tidyverse)

df %>%
  mutate_all(as.character) %>%
  mutate(uniquecounts = map2_dbl(names, partners, 
                       ~ n_distinct(c(.x, str_split(.y, ", ")[[1]]))))


#    names                    partners uniquecounts
#1    John  Mary, Ashley, John, Kate            4
#2    Mary Charlie, John, Mary, John            3
#3 Charlie               Kate, Marcy            3
#4   David              Mary, Claire            3

---

With same logic in base R

df[] <- lapply(df, as.character)
as.numeric(mapply(function(x, y) length(unique(c(x, y))), 
          df$names, strsplit(df$partners, ", ")))
#[1] 4 3 3 3