sum() condition in ifelse statement

user_13321647 0
Viewed 142

This question is related to this question My question is about R: How to number each repetition in a table in R?

Where basically the repetitions are numbered. E.g two repetitions: 1,2 ; three repetitions: 1,2,3 etc... But if the value is unique (only one time) it should have not 1 but NA

data: (from akrun, many thanks!)

df1 <- structure(list(Fullname = c("Peter", "Peter", "Alison", "Warren", 
                                   "Jack", "Jack", "Jack", "Jack", "Susan", "Susan", "Henry", "Walison", 
                                   "Tinder", "Peter", "Henry", "Tinder")), row.names = c(NA, -16L
                                   ), class = "data.frame")

my solution would be this:

df1 %>% 
  group_by(Fullname) %>% 
  mutate(newcol = seq_along(Fullname)) 

  Fullname newcol
   <chr>     <int>
 1 Peter         1
 2 Peter         2
 3 Alison        1
 4 Warren        1
 5 Jack          1
 6 Jack          2
 7 Jack          3
 8 Jack          4
 9 Susan         1
10 Susan         2
11 Henry         1
12 Walison       1
13 Tinder        1
14 Peter         3
15 Henry         2
16 Tinder        2

Now I try to set each value that occurs once (e.g. Alison, Warren and Henry) to NAlike akrun did here My question is about R: How to number each repetition in a table in R?

My code is with a ifelse statement checking if the sum of the group is >1.

df1 %>% 
  group_by(Fullname) %>% 
  mutate(newcol = seq_along(Fullname)) %>% 
  mutate(newcol = ifelse(sum(newcol)>1, newcol, NA))

but I get:

 Fullname newcol
   <chr>     <int>
 1 Peter         1
 2 Peter         1
 3 Alison       NA
 4 Warren       NA
 5 Jack          1
 6 Jack          1
 7 Jack          1
 8 Jack          1
 9 Susan         1
10 Susan         1
11 Henry         1
12 Walison      NA
13 Tinder        1
14 Peter         1
15 Henry         1
16 Tinder        1

And I can't grasp why?

dplyr if-statement r sum
2 Answers

We need if/else here instead of ifelse as ifelse requires all arguments to be same length, sum returns a single value and if it is TRUE, then all becomes TRUE

library(dplyr)
df1 %>% 
  group_by(Fullname) %>% 
  mutate(newcol = row_number(), 
       newcol = if(sum(newcol)> 1) newcol else NA) %>%
  ungroup

-output

# A tibble: 16 × 2
   Fullname newcol
   <chr>     <int>
 1 Peter         1
 2 Peter         2
 3 Alison       NA
 4 Warren       NA
 5 Jack          1
 6 Jack          2
 7 Jack          3
 8 Jack          4
 9 Susan         1
10 Susan         2
11 Henry         1
12 Walison      NA
13 Tinder        1
14 Peter         3
15 Henry         2
16 Tinder        2

Now, we look at the issue. The 'newcol2' values are recycled values of single TRUE/FALSE. In the ifelse, as all arguments need to be same length, the logical part is just of length 1.

df1 %>% 
   group_by(Fullname) %>% 
   mutate(newcol = row_number(), newcol2 = sum(newcol) > 1)
# A tibble: 16 × 3
# Groups:   Fullname [8]
   Fullname newcol newcol2
   <chr>     <int> <lgl>  
 1 Peter         1 TRUE   
 2 Peter         2 TRUE   
 3 Alison        1 FALSE  
 4 Warren        1 FALSE  
 5 Jack          1 TRUE   
 6 Jack          2 TRUE   
 7 Jack          3 TRUE   
 8 Jack          4 TRUE   
 9 Susan         1 TRUE   
10 Susan         2 TRUE   
11 Henry         1 TRUE   
12 Walison       1 FALSE  
13 Tinder        1 TRUE   
14 Peter         3 TRUE   
15 Henry         2 TRUE   
16 Tinder        2 TRUE

The way to tackle is replicate to make the lengths same

df1 %>% 
  group_by(Fullname) %>% 
  mutate(newcol = seq_along(Fullname)) %>% 
  mutate(newcol = ifelse(rep(sum(newcol)>1, n()), newcol, NA))
# A tibble: 16 × 2
# Groups:   Fullname [8]
   Fullname newcol
   <chr>     <int>
 1 Peter         1
 2 Peter         2
 3 Alison       NA
 4 Warren       NA
 5 Jack          1
 6 Jack          2
 7 Jack          3
 8 Jack          4
 9 Susan         1
10 Susan         2
11 Henry         1
12 Walison      NA
13 Tinder        1
14 Peter         3
15 Henry         2
16 Tinder        2

In order to understand it better, just take a simple vector

> v1 <- c(1:5)
> sum(v1) > 4
[1] TRUE
> ifelse(sum(v1) > 4, v1, NA)
[1] 1

The sum here is 15 and it is definitely greater than 4. As soon as the TRUE is found, it just returns the first element of the vector i.e. 1 and stops. In the %>% also, this is what is happening, but because there is recycling, the 1 gets repeated to fill the whole group

user_37322710

Provided you have already understood why the error occurs. I however, think sum is not the right function to use here. Yes, you can make it work but it seems convoluted. You can use n() to get number of rows in the group and check it with an if condition.

library(dplyr)

df1 %>%
  group_by(Fullname) %>%
  mutate(newcol = if(n() == 1) NA else row_number()) %>%
  ungroup

# Fullname newcol
#   <chr>     <int>
# 1 Peter         1
# 2 Peter         2
# 3 Alison       NA
# 4 Warren       NA
# 5 Jack          1
# 6 Jack          2
# 7 Jack          3
# 8 Jack          4
# 9 Susan         1
#10 Susan         2
#11 Henry         1
#12 Walison      NA
#13 Tinder        1
#14 Peter         3
#15 Henry         2
#16 Tinder        2
user_39629140
Related