dplyr group_by分组后用grepl获取参与者首次切换偏好的日期与研究日

首先定义原始数据框：

df <- data.frame(participant = c(1,1,1,2,2,2,2),
               food = c("pizza", "turkey", "turkey", "pizza", "pizza", "pizza", "turkey"),
               date = c("2012-01-01", "2012-01-02", "2012-01-03","2012-01-01", "2012-01-02", "2012-01-03", "2012-01-04"))

需求说明 #

需要为每个参与者新增两个变量：

• study_day：首次切换偏好至turkey的研究日（即该记录在对应参与者分组内的行号）
• change_date：首次切换偏好至turkey对应的日期

期望的结果数据框如下：

df2 <- data.frame(participant = c(1,1,1,2,2,2,2),
                  food = c("pizza", "turkey", "turkey", "pizza", "pizza", "pizza", "turkey"),
                  date = c("2012-01-01", "2012-01-02", "2012-01-03","2012-01-01", "2012-01-02", "2012-01-03", "2012-01-04"),
                 study_day = c(2,2,2,4,4,4,4),
                 change_date = c("2012-01-02","2012-01-02","2012-01-02","2012-01-04","2012-01-04","2012-01-04","2012-01-04"))

错误原因 #

之前的代码中使用了df$food和df$date，这会直接调用整个原始数据框的列，完全忽略了group_by(participant)的分组逻辑，导致结果是全局范围内第一个turkey的记录，而非每个参与者分组内的结果。

正确解决方案 #

去掉df$前缀，直接引用分组内的变量即可，以下提供两种可行写法：

写法一：使用which()定位

library(dplyr)

df %>% 
  group_by(participant) %>% 
  mutate(
    study_day = which(food == "turkey")[1],
    change_date = date[which(food == "turkey")[1]]
  ) %>% 
  ungroup() # 可选，用于取消分组状态

写法二：使用first()和row_number()简化

library(dplyr)

df %>% 
  group_by(participant) %>% 
  mutate(
    study_day = row_number()[food == "turkey"][1],
    change_date = first(date[food == "turkey"])
  ) %>% 
  ungroup()

两种写法均能生成符合预期的结果，确保每个参与者的study_day和change_date对应自身首次切换至turkey的记录。

内容的提问来源于stack exchange，提问作者Kevin Li