如何从R语言DataFrame宽格式数据生成药物类别指示变量？

需求说明 #

现有宽格式DataFrame，记录受试者多次访视的药物服用数据（DRUG1-DRUG3），需生成三个指示变量：

• CLASS1：该访视是否服用A/B/C类药物（是=1，否=0）
• CLASS2：该访视是否服用D/E/F类药物（是=1，否=0）
• CLASS3：该访视是否服用G/H/I类药物（是=1，否=0）

示例数据：

ID <- c(1,1,2,2,3,3,4,4,4)
Visit <- c(1,2,1,2,1,2,1,2,3)
DRUG1 <- c("A","A","A","A","G","G","D","D","G")
DRUG2 <- c("A","G","B","B","G","G","D","D","G")
DRUG3 <- c("A","G","B","B","G","G","D","D","G")
df <- data.frame(ID, Visit, DRUG1, DRUG2, DRUG3)

期望输出的指示变量：

• CLASS1: (1,1,1,1,0,0,0,0,0)
• CLASS2: (0,0,0,0,0,0,1,1,0)
• CLASS3: (0,1,0,0,1,1,0,0,1)

方法一：Base R实现（无需额外包） #

先定义药物分类映射，再逐行检查是否包含对应类别药物：

# 定义药物分类
drug_classes <- list(
  CLASS1 = c("A", "B", "C"),
  CLASS2 = c("D", "E", "F"),
  CLASS3 = c("G", "H", "I")
)

# 生成指示变量
df$CLASS1 <- apply(df[, grep("DRUG", colnames(df))], 1, function(x) any(x %in% drug_classes$CLASS1) + 0)
df$CLASS2 <- apply(df[, grep("DRUG", colnames(df))], 1, function(x) any(x %in% drug_classes$CLASS2) + 0)
df$CLASS3 <- apply(df[, grep("DRUG", colnames(df))], 1, function(x) any(x %in% drug_classes$CLASS3) + 0)

方法二：Tidyverse实现（dplyr） #

利用rowwise()和c_across()逐行处理，代码更简洁：

library(dplyr)

df <- df %>%
  rowwise() %>%
  mutate(
    CLASS1 = +any(c_across(starts_with("DRUG")) %in% c("A", "B", "C")),
    CLASS2 = +any(c_across(starts_with("DRUG")) %in% c("D", "E", "F")),
    CLASS3 = +any(c_across(starts_with("DRUG")) %in% c("G", "H", "I"))
  ) %>%
  ungroup()

验证结果 #

运行上述代码后，查看df的CLASS1-CLASS3列，即可得到符合预期的指示变量。

内容的提问来源于stack exchange，提问作者Jenn0804