CODEWITHSUNDEEP
rdplyrtidyverse
Chris
Chris

Reputation: 2081

sapply(), with unique() and conditional - R

I'm trying to do a simple and straightforward output to see (1) the number of distinct values per variable, and (2) show which distinct values are those for the variables with less than X distinct values. When you run this:

sapply(mtcars, function(z) NROW(unique(z)))

It gives a very simple, straightforward information per variable:

 mpg 25
 cyl 3
disp 27
  hp 22
drat 22
  wt 29
qsec 30
  vs 2
  am 2
gear 3
carb 6

However, I still need to show the distinct values. Let's say we will show the distinct values for the variables with less than 10 distinct values. I have achieved this so far:

sapply(mtcars, function(z) if(NROW(unique(z)) < 10) {paste0(NROW(unique(z)), " ; ", unique(z))} else {NROW(unique(z))})

And it shows a messy summary. I'm looking for something like this:

 mpg 25
 cyl 3 ; 6 4 8
disp 27
  hp 22
drat 22
  wt 29
qsec 30
  vs 2 ; 0 1
  am 2 ; 1 0
gear 3 ; 4 3 5
carb 6 ; 4 1 2 3 6 8

Upvotes: 1

Views: 780

Answers (3)

jazzurro
jazzurro

Reputation: 23574

Here is an option that I came up with. First I created a data frame containing the number of unique values in each variable, which is tmp1. Then, I created a character vector containing unique values in each variable. Since you specified that you want to print out unique values if there are less than 10 unique values, I handled that in the if_else() part. Then, I bound tmp1 and tmp2, and changed the order of columns as well as a variable name.

library(dplyr)

summarize_all(mtcars,
              .funs = list(~n_distinct(.))) %>% 
  stack -> tmp1 

summarize_all(mtcars,
              .funs = list(~if_else(n_distinct(.) < 10,
                                    toString(unique(.)),
                                    "More than 10 unique values"))) %>% 
  unlist -> tmp2

bind_cols(tmp1, distinct_value = tmp2) %>%
  select(variable = ind, everything())

#   variable values             distinct_value
#1       mpg     25 More than 10 unique values
#2       cyl      3                    6, 4, 8
#3      disp     27 More than 10 unique values
#4        hp     22 More than 10 unique values
#5      drat     22 More than 10 unique values
#6        wt     29 More than 10 unique values
#7      qsec     30 More than 10 unique values
#8        vs      2                       0, 1
#9        am      2                       1, 0
#10     gear      3                    4, 3, 5
#11     carb      6           4, 1, 2, 3, 6, 8

Upvotes: 2

Chris
Chris

Reputation: 2081

I took @H1 answer as it's exactly the output expected (a simple one):

sapply(mtcars, function(z) if(length(unique(z)) < 10) {paste0(length(unique(z)), "; ", toString(unique(z)))} else {length(unique(z))})

Upvotes: 0

Ronak Shah
Ronak Shah

Reputation: 389335

You can try something like this and create nested list conditional on length of unique values in each column.

sapply(mtcars, function(x) {
   uniq <- unique(x) 
   if (length(uniq) < 10)
     list(no_uniq_values = length(uniq), uniq_values = uniq)
   else
     length(uniq)
})

#$mpg
#[1] 25

#$cyl
#$cyl$no_uniq_values
#[1] 3

#$cyl$uniq_values
#[1] 6 4 8


#$disp
#[1] 27

#$hp
#[1] 22

#$drat
#[1] 22
#.....
#.....

Upvotes: 1

Related Questions