R语言：按分组批量统计多列中'yes'的占比

需求说明 #

按group列分组，统计数据框中其余每一列里"yes"出现的占比（忽略NA值，结果保留两位小数），示例数据及期望结果如下：

示例数据：

df <- data.frame(group = c("a","a","a","a","b","b","b","b"),
                 female = c("yes","no","yes","no","yes","yes","yes","no"),
                 alcohol = c("yes","no",NA,"no","yes","yes","no","no"))

期望结果：

> df_result
  group female alcohol
1     a   0.50    0.33
2     b   0.75    0.50

解决方案 #

使用tidyverse中的dplyr包，通过group_by()+summarise()+across()组合实现，无需逐一指定列名：

library(tidyverse)

# 示例数据
df <- data.frame(group = c("a","a","a","a","b","b","b","b"),
                 female = c("yes","no","yes","no","yes","yes","yes","no"),
                 alcohol = c("yes","no",NA,"no","yes","yes","no","no"))

# 核心代码
df_result <- df %>%
  group_by(group) %>%
  # 对所有非分组列计算yes的占比，忽略NA
  summarise(across(everything(), ~ mean(.x == "yes", na.rm = TRUE))) %>%
  ungroup() %>%
  # 保留两位小数，匹配示例格式
  mutate(across(-group, ~ round(.x, 2)))

# 查看结果
df_result

代码解释 #

• group_by(group)：按group列进行分组
• across(everything(), ~ mean(.x == "yes", na.rm = TRUE))：
  • across(everything())：自动遍历除分组列外的所有列，无需手动列名
  • .x == "yes"：将列中每个元素转为逻辑值（等于"yes"为TRUE，否则FALSE）
  • mean(..., na.rm = TRUE)：逻辑值的均值即为TRUE的占比，na.rm = TRUE忽略NA值（和示例中a组alcohol的计算逻辑一致：3个非NA值里1个yes，占比1/3≈0.33）
• ungroup()：取消分组状态，避免后续操作受分组影响
• mutate(across(-group, ~ round(.x, 2)))：对除group外的列保留两位小数，匹配期望结果格式

扩展说明 #

如果只想统计特定类型的列（比如仅字符型列），可以把across(everything())改成：

across(where(is.character), ~ mean(.x == "yes", na.rm = TRUE))

如果需要排除某几列，比如排除group和另一列xxx，可以用：

across(-c(group, xxx), ~ mean(.x == "yes", na.rm = TRUE))

内容的提问来源于stack exchange，提问作者BPeif