在 r 中检查逻辑变量之和是否大于 n，使用 na

Question

I have a dataframe with 5 binary variables ( TRUE or FALSE , but represented as 0 or 1 for convenience) which can have missing values:

df <- data.frame(a = c(1,0,1,0,0,...),
                 b = c(1,0,NA,0,1,...),
                 c = c(1,0,1,0,NA,...),
                 d = c(0,1,1,NA,NA,...),
                 e = c(0,0,0,1,1,...))

     a  b  c  d  e
 1   1  1  1  0  0
 2   0  0  0  1  0
 3   1 NA  1  1  0
 4   0  0  0 NA  1
 5   0  1 NA NA  1
...

Now I want to make a variable that indicates whether the observation satisfies more than two conditions out of the five, that is, whether the sum of a , b , c , d , and e is greater than 2.

For the first row and the second row, the values are obviously TRUE and FALSE respectively. For the third row, the value should be TRUE , since the sum is greater than 2 regardless of whether b is TRUE or FALSE . For the third row, the value should be FALSE , since the sum is less than or equal to 2 regardless of whether d is TRUE or FALSE . For the fifth row, the value should be NA , since the sum can range from 2 to 4 depending on c and d . So the desirable vector is c(TRUE, FALSE, TRUE, FALSE, NA, ...) .

Here is my attempt:

df %>%
  mutate(a0 = ifelse(is.na(a), 0, a),
         b0 = ifelse(is.na(b), 0, b),
         c0 = ifelse(is.na(c), 0, c),
         d0 = ifelse(is.na(d), 0, d),
         e0 = ifelse(is.na(e), 0, e),
         a1 = ifelse(is.na(a), 1, a),
         b1 = ifelse(is.na(b), 1, b),
         c1 = ifelse(is.na(c), 1, c),
         d1 = ifelse(is.na(d), 1, d),
         e1 = ifelse(is.na(e), 1, e)
         ) %>%
  mutate(summin = a0 + b0 + c0 + d0 + e0,
         summax = a1 + b1 + c1 + d1 + e1) %>%
  mutate(f = ifelse(summax <= 2,
                    FALSE,
                    ifelse(summin >= 3, TRUE, NA)))

This did work, but I had to make too many redunant variables, plus the code would be too lengthy if there were more variables. Is there any better solution?

Answer 1

I just noticed that you want NA in case the outcome of the missing value will determine the TRUE/FALSE outcome, so I have changed the answer.

Combining two if_else statements can first test if the row already have a sum of more than 2, and if not, check if the row sum plus the number of missing values is 2 or less.

library(tidyverse)
n <- 2
want <- ifelse(rowSums(df, na.rm = TRUE) > n, 
               TRUE, 
               if_else((rowSums(df, na.rm = TRUE) + rowSums(is.na(df)))<=n,
                        FALSE, 
                        NA))

If you want to stick to base-R you can use the function ifelse() instead.

Answer 2

I am not sure what you mean by "For the fifth row, the value should be NA, since the sum can range from 2 to 4 depending on c and d."

But the following results in the vector you wish for:

test <- ifelse(is.na(df$c), NA, ifelse(rowSums(df[1:5,], na.rm=T) > 2, TRUE, FALSE))

If there is an NA value in the column c, an NA value will be inserted in the new vector test . Else, it is tested if the sum of the first 5 columns is greater than 2 - if true, TRUE will be inserted and FALSE when the sum is lower than or exactly two.