基于多行条件的新列答案

【问题标题】：New column based on multiple line conditions基于多行条件的新列
【发布时间】：2021-11-07 07:10:34
【问题描述】：

我有这个 df

data.frame(Name = c("AI147", "AI147", "AI147", "AI147", "AI147", 
                         "AI20", "AI20", "AI87", "AI88", "AI88", "AI88", "AI65", "AI65"),
                         Presence1 = c("both_type1", "soil", "soil", "water", "both_type2", 
                         "soil", "water", "both_type2", "soil", "soil", "soil", "water", 
                         "water"))

我想根据每个名称的数据创建一个条件列（最终）。 (1) 如果给定名称具有多于 1 种存在类型，或仅存在“both_type1”或“both_type2”，则最终 = 两者，(2) 如果给定名称仅存在“土壤”，则最终 = 土壤， (3) 如果给定的 Name 仅存在“water”，则 Final = water，因此表格看起来像这样

data.frame(Name = c("AI147", "AI147", "AI147", "AI147", "AI147", 
                         "AI20", "AI20", "AI87", "AI88", "AI88", "AI88", "AI65", "AI65"),
                         Presence1 = c("both_type1", "soil", "soil", "water", "both_type2", 
                         "soil", "water", "both_type2", "soil", "soil", "soil", "water", 
                         "water"),
                         Final = c("both", "both", "both", "both", "both", "both", 
                         "both", "both", "soil", "soil", "soil", "water", "water"))

我已经尝试了几种我在网站上找到的方法，但没有一种方法能做到这一点。

【问题讨论】：

标签： r multiple-conditions

【解决方案1】：

我们可以使用n_distinct 或使用str_detect 创建条件，即如果“Presence1”包含any“both”子字符串或具有多个唯一值（n_distinct），则在按“名称”分组后，然后return 'both' 否则返回 'Presence1' 的值

library(dplyr)
library(stringr)
df1 <- df %>%
    group_by(Name) %>% 
    mutate(Final = case_when(any(str_detect(Presence1, 
       "both")|n_distinct(Presence1) > 1) ~ 'both',
        TRUE ~ Presence1 )) %>%
    ungroup

-输出

df1
# A tibble: 13 x 3
   Name  Presence1  Final
   <chr> <chr>      <chr>
 1 AI147 both_type1 both 
 2 AI147 soil       both 
 3 AI147 soil       both 
 4 AI147 water      both 
 5 AI147 both_type2 both 
 6 AI20  soil       both 
 7 AI20  water      both 
 8 AI87  both_type2 both 
 9 AI88  soil       soil 
10 AI88  soil       soil 
11 AI88  soil       soil 
12 AI65  water      water
13 AI65  water      water

【讨论】：