取消组合变量(combn)作为向量的嵌套

Question

使用下面的代码，我设法获得了一个很好的combination：

tibble(
    x = list(c(1, 2, 3), c(4,5,6))
  ) %>% 
  mutate(
      combination = 
        x %>% 
        map(
            .f = combn
          , 2
          ) %>% 
        map(.f = t)
    ) %>% 
  unnest(combination)

# A tibble: 6 x 2
  x         combination[,1]  [,2]
  <list>              <dbl> <dbl>
1 <dbl [3]>               1     2
2 <dbl [3]>               1     3
3 <dbl [3]>               2     3
4 <dbl [3]>               4     5
5 <dbl [3]>               4     6
6 <dbl [3]>               5     6

然而，当使用View()函数观察时，我得到：

​

​

如何继续将combination显示为向量？即：

​

​

Answer 1

我们可以在combn中指定simplify = FALSE来返回list，而不是强制matrix

library(purrr)
library(dplyr)
library(tidyr)
tbl1 <- tibble(
   x = list(c(1, 2, 3), c(4,5,6))
  ) %>% 
 mutate(
  combination = 
    x %>% 
    map(
        .f = combn
      , 2, simplify = FALSE
      )) 

现在，执行unnest

out <- tbl1 %>%
   unnest(combination)

out
# A tibble: 6 x 2
#  x         combination
#  <list>    <list>     
#1 <dbl [3]> <dbl [2]>  
#2 <dbl [3]> <dbl [2]>  
#3 <dbl [3]> <dbl [2]>  
#4 <dbl [3]> <dbl [2]>  
#5 <dbl [3]> <dbl [2]>  
#6 <dbl [3]> <dbl [2]>  

检查View

​

​

Answer 2

下面是一个可能有帮助的data.table选项

library(data.table)
library(tidyr)
unnest(setDT(df)[, combination := lapply(x, function(v) combn(v, 2, simplify = FALSE))], combination)