基于多个条件创建新变量答案

【问题标题】：Creating new variable based on more than one condition基于多个条件创建新变量
【发布时间】：2021-06-15 15:47:41
【问题描述】：

我正在尝试根据某些条件创建一个新变量。我有以下数据：

df <- data.frame(ID = c("A1","A1","A2","A2","A3","A4","A4"),
                 type = c("small","large","small","large","large","small","large"),
                 code = c("B9", "[0,20]","B9","[20,40]","[0,20]","B9","[40,60]" ))

这给出了：

    ID  type   code
1   A1  small   B9
2   A1  large   [0,20]
3   A2  small   B9
4   A2  large   [20,40]
5   A3  large   [0,20]
6   A4  small   B9
7   A4  large   [40,60]

我想创建一个新变量（code2），它基于 type == large 和 code 的相应值，同时按 ID。所以 ID - A1 应该有 [0,20] 作为它的代码 2。我想实现以下目标：

    ID  type   code       code2
1   A1  small   B9        [0,20]    
2   A1  large   [0,20]    [0,20] 
3   A2  small   B9        [20,40]
4   A2  large   [20,40]   [20,40]
5   A3  large   [0,20]    [0,20] 
6   A4  small   B9        [40,60]
7   A4  large   [40,60]   [40,60]

据我所知，我正在尝试使用dplyr 和ifelse，但没有运气。

【问题讨论】：

标签： r

【解决方案1】：

我们可以在dplyr 中使用分组操作，即按“ID”分组，提取“类型”值为“大”的“代码”（假设每个“类型”中没有重复值）身份证'

library(dplyr)
df <- df %>% 
   group_by(ID) %>%
   mutate(code2 = code[type == 'large']) %>%
   ungroup

-输出

df
# A tibble: 7 x 4
  ID    type  code    code2  
  <chr> <chr> <chr>   <chr>  
1 A1    small B9      [0,20] 
2 A1    large [0,20]  [0,20] 
3 A2    small B9      [20,40]
4 A2    large [20,40] [20,40]
5 A3    large [0,20]  [0,20] 
6 A4    small B9      [40,60]
7 A4    large [40,60] [40,60]

如果有重复，使用match，它会给出第一个匹配索引的索引

df <- df %>%
       group_by(ID) %>%
       mutate(code2 = code[match('large', type)]) %>%
       ungroup

【讨论】：

谢谢。如果每个 ID 中的“类型”值重复，我该如何使用？
谢谢。我尝试了match 并得到 [0,20] 对所有行重复。它应该是这样工作的吗？
@cajt 你用过group_by(ID)
我用了group_by(ID)，但没有用。
@cajt 我在您的示例中尝试了该方法。它提供与第一个解决方案相同的输出

【解决方案2】：

data.table 选项

> setDT(df)[, code2 := code[type == "large"], ID][]
   ID  type    code   code2
1: A1 small      B9  [0,20]
2: A1 large  [0,20]  [0,20]
3: A2 small      B9 [20,40]
4: A2 large [20,40] [20,40]
5: A3 large  [0,20]  [0,20]
6: A4 small      B9 [40,60]
7: A4 large [40,60] [40,60]

【讨论】：