Tidy way of NA management by groups?

Question

my issue is the following:

I have a tibble which I want to modify with 3 different cases:

All values in group are NA
At least one but not all are NA. In this case replace NA with an arbitrary value (ex: 0.5)
None are NA

Example: (with group_by ind)

    a1 = c(0.3,0.1,NA,0.7,0.2)
    a2 = rep(NA,5)
    a3 = c(0.1,0.3,0.5,0.7,0.8)
    tibble(ind = c(rep("A",5),rep("B",5),rep("C",5)),
       value = c(a1,a2,a3)

segment of group A should yield c(0.3,0.1,0.5,0.7,0.2)

segment of group B should yield rep(NA,5)

segment of group C should stay the same

I've tried with if, ifelse and case_when statements but I think I'm missing something very obvious. All help is appreciated.

Correct, however if the whole group is `NA` or none of them are keep all `NA`s or keep all values. — GonzaloXavier, Jul 27 '20 at 21:14

Matt · Accepted Answer · 2020-07-27T21:37:33.793

Edit:

Here's a hack way to do it, though I know there is a more concise way:

library(dplyr)
df %>% 
  group_by(ind) %>% 
  mutate_if(is.logical, as.numeric) %>% 
  mutate(a1 = case_when(is.na(a1) & sum(is.na(a1)) < length(a1) ~ 0.5, TRUE ~ a1),
         a2 = case_when(is.na(a2) & sum(is.na(a2)) < length(a2) ~ 0.5, TRUE ~ a2),
         a3 = case_when(is.na(a3) & sum(is.na(a3)) < length(a3) ~ 0.5, TRUE ~ a3))

Edit2: Here's the more concise way

point_five <- function(x){
  x = case_when(is.na(x) & sum(is.na(x)) < length(x) ~ 0.5, TRUE ~ x)
}

df %>%
  group_by(ind) %>% 
  mutate_if(is.logical, as.numeric) %>% 
  mutate(across(.cols = c(a1:a3), ~ point_five(.)))

This gives us:

# A tibble: 5 x 4
# Groups:   ind [1]
  ind      a1    a2    a3
  <chr> <dbl> <dbl> <dbl>
1 A       0.3    NA   0.1
2 A       0.1    NA   0.3
3 A       0.5    NA   0.5
4 A       0.7    NA   0.7
5 A       0.2    NA   0.8

If we have df2, containing two groups for ind, group_by will give us:

   ind      a1    a2    a3
   <chr> <dbl> <dbl> <dbl>
 1 A       0.3    NA   0.1
 2 A       0.5    NA   0.3
 3 A       0.5    NA   0.5
 4 A       0.7    NA   0.7
 5 A       0.2    NA   0.8
 6 B       0.5    NA   0.1
 7 B       0.5    NA   0.3
 8 B       0.5    NA   0.5
 9 B       0.5    NA   0.7
10 B       0.2    NA   0.8

df2

structure(list(ind = c("A", "A", "A", "A", "A", "B", "B", "B", 
"B", "B"), a1 = c(0.3, 0.5, NA, 0.7, 0.2, NA, 0.5, NA, NA, 0.2
), a2 = c(NA, NA, NA, NA, NA, NA, NA, NA, NA, NA), a3 = c(0.1, 
0.3, 0.5, 0.7, 0.8, 0.1, 0.3, 0.5, 0.7, 0.8)), row.names = c(NA, 
-10L), class = c("tbl_df", "tbl", "data.frame"))

I may have misexpressed myself. The 0.5 is not becuase it happens to be the last value, its an arbitrary number. Will edit for clarity. — GonzaloXavier, Jul 27 '20 at 21:10

score 1 · Answer 2 · answered Jul 27 '20 at 21:48

An approach using case_when

df %>%
  group_by(ind) %>%
  mutate(value2 = case_when(
    !all(is.na(value)) & is.na(value) ~ 0.5,
    TRUE ~ value
  ))


#-----

# A tibble: 15 x 3
# Groups:   ind [3]
   ind   value value2
   <chr> <dbl>  <dbl>
 1 A       0.3    0.3
 2 A       0.1    0.1
 3 A      NA      0.5
 4 A       0.7    0.7
 5 A       0.2    0.2
 6 B      NA     NA  
 7 B      NA     NA  
 8 B      NA     NA  
 9 B      NA     NA  
10 B      NA     NA  
11 C       0.1    0.1
12 C       0.3    0.3
13 C       0.5    0.5
14 C       0.7    0.7
15 C       0.8    0.8

Example data

a1 = c(0.3,0.1,NA,0.7,0.2)
a2 = rep(NA,5)
a3 = c(0.1,0.3,0.5,0.7,0.8)
df <- tibble(ind = c(rep("A",5),rep("B",5),rep("C",5)),
             value = c(a1,a2,a3))

Tidy way of NA management by groups?

2 Answers2