我们可以首先通过查找定界符;的出现次数max来确定需要多少列。然后paste使用“country”字符串在separate的into =参数中获取信息。

library(tidyverse)

col_number <- max(str_count(start_problem$country, ";") + 1)

start_problem %>% separate(country, 
                           into = paste0("country", seq_len(col_number)), 
                           sep = ";")

  continent country1 country2 country3
1    Europe   France  Germany    Italy
2      Asia    Japan     <NA>     <NA>

另一种选择是先用separate_rows分隔行。创建一个列，其中包含用于pivot_wider的名称，以使数据更宽，如下所示：

library(tidyverse)
start_problem %>%
  separate_rows(country, sep = ";") %>%
  mutate(col_name = paste0("country", row_number()), .by = continent) %>%
  pivot_wider(names_from = col_name, values_from = country)
#> # A tibble: 2 × 4
#>   continent country1 country2 country3
#>   <chr>     <chr>    <chr>    <chr>   
#> 1 Europe    France   Germany  Italy   
#> 2 Asia      Japan    <NA>     <NA>

创建于2023-03-31带有reprex v2.0.2

碱基R：

cbind(start_problem[1], read.csv2(text=start_problem[,2], header = FALSE))
  continent     V1      V2    V3
1    Europe France Germany Italy
2      Asia  Japan

如果您严格需要NA，则使用

cbind(start_problem[1], read.csv2(text=start_problem[,2], header = FALSE, na.strings = ''))
  continent     V1      V2    V3
1    Europe France Germany Italy
2      Asia  Japan    <NA>  <NA>