如何在多列上执行IFELSE？

Question

我有示例dataframe

dat <- data.frame(
  variable1 = c(NA,2,3,4,5,6,99),
  variable2 = c(NA,2,99,4,5,6,7),
  variable3 = c(NA,2,3,4,5,6,7),
  variable4 = c(5:11),
  variable5 = c(1,2,3,4,5,6,999),
  variable6 = c(1,2,3,4,999,6,7),
  variable7 = c(1:7)
)

  variable1 variable2 variable3 variable4 variable5 variable6 variable7 
1        NA        NA        NA         5         1         1         1 
2         2         2         2         6         2         2         2 
3         3        99         3         7         3         3         3 
4         4         4         4         8         4         4         4
5         5         5         5         9         5       999         5
6         6         6         6        10         6         6         6
7        99         7         7        11       999         7         7 

我想执行一个ifelse，说明是否变量1和变量2和变量3是na， 采用变量4，变量5，变量6， 否则，请变量1，变量2，变量3 到新列 变量8，变量9，变量10。

因此，新的数据框应该看起来像这样

  variable1 variable2 variable3 variable4 variable5 variable6 variable7 variable8 variable9 variable10
1        NA        NA        NA         5         1         1         1         5         1         1
2         2         2         2         6         2         2         2         2         2         2 
3         3        99         3         7         3         3         3         3        99         3
4         4         4         4         8         4         4         4         4         4         4 
5         5         5         5         9         5       999         5         5         5         5
6         6         6         6        10         6         6         6         6         6         6
7        99         7         7        11       999         7         7         99        7         7

我更喜欢Dplyr解决方案；）

Answer 1

dplyr :: case_when是完美的：

library(dplyr)
dat %>% 
  mutate(across(all_of(1:7), ~as.numeric(.))) %>% #important for all data to be numeric
  mutate(variable8 = case_when(is.na(variable1) & is.na(variable2) & is.na(variable3) ~ variable4,
                               TRUE ~ variable1)) %>%
  mutate(variable9 = case_when(is.na(variable1) & is.na(variable2) & is.na(variable3) ~ variable5,
                               TRUE ~ variable2)) %>%
  mutate(variable10 = case_when(is.na(variable1) & is.na(variable2) & is.na(variable3) ~ variable6,
                               TRUE ~ variable3))

  variable1 variable2 variable3 variable4 variable5
1        NA        NA        NA         5         1
2         2         2         2         6         2
3         3        99         3         7         3
4         4         4         4         8         4
5         5         5         5         9         5
6         6         6         6        10         6
7        99         7         7        11       999
  variable6 variable7 variable8 variable9 variable10
1         1         1         5         1          1
2         2         2         2         2          2
3         3         3         3        99          3
4         4         4         4         4          4
5       999         5         5         5          5
6         6         6         6         6          6
7         7         7        99         7          7

Answer 2

另一个可能的解决方案：

library(tidyverse)

dat %>% 
  mutate(pmap_dfr(., ~ (if (all(is.na(c(...)[1:3]))) {c(...)[4:6]} else 
    {c(...)[1:3]}) %>% set_names(str_c("variable", 8:10))))

#>   variable1 variable2 variable3 variable4 variable5 variable6 variable7
#> 1        NA        NA        NA         5         1         1         1
#> 2         2         2         2         6         2         2         2
#> 3         3        99         3         7         3         3         3
#> 4         4         4         4         8         4         4         4
#> 5         5         5         5         9         5       999         5
#> 6         6         6         6        10         6         6         6
#> 7        99         7         7        11       999         7         7
#>   variable8 variable9 variable10
#> 1         5         1          1
#> 2         2         2          2
#> 3         3        99          3
#> 4         4         4          4
#> 5         5         5          5
#> 6         6         6          6
#> 7        99         7          7