R语言如何批量对每列按自身分组计数并计算占比生成统一数据框

最优实现方案（tidyverse 生态） #

使用purrr包的列遍历功能直接批量处理所有列，自动合并为统一结果数据框，无需手动维护循环逻辑：

library(tidyverse)

# 示例数据
a<- rep(1:5, 5) 
b <- rep(1:5, 5) 
c <- rep(1:5, 5) 
d <- rep(1:5, 5) 
df <- data.frame(a=a, b=b, c=c, d=d)

# 批量处理代码
all_col_result <- map_dfr(names(df), function(col_name) {
  df %>%
    group_by(across(all_of(col_name))) %>%
    summarise(count = n()) %>%
    pivot_wider(names_from = all_of(col_name), values_from = count) %>%
    mutate(
      sum = rowSums(across(where(is.numeric))),
      neg = (`1` + `2`)/sum,
      pos = (`4` + `5`)/sum,
      neut = `3`/sum
    ) %>%
    select(pos, neg, neut) %>%
    # 可选：新增字段标记当前结果对应的原始列名
    mutate(source_col = col_name, .before = 1)
})

# 输出结果
all_col_result

修正后的for循环实现 #

如果你更习惯用for循环，修改你的错误写法后可用版本如下：

# 初始化空结果容器
all_col_result <- data.frame()

for (col_name in names(df)) {
  temp_res <- df %>%
    group_by(across(all_of(col_name))) %>%
    summarise(count = n()) %>%
    pivot_wider(names_from = all_of(col_name), values_from = count) %>%
    mutate(
      sum = rowSums(across(where(is.numeric))),
      neg = (`1` + `2`)/sum,
      pos = (`4` + `5`)/sum,
      neut = `3`/sum
    ) %>%
    select(pos, neg, neut) %>%
    mutate(source_col = col_name, .before = 1)
  # 拼接单例结果到总表
  all_col_result <- rbind(all_col_result, temp_res)
}

说明 #

• 原代码报错核心原因是!!df[i,]用法错误，group_by需要传入列名而非列的取值，用across(all_of(列名字符串))可以实现动态指定分组列
• 用rowSums(across(where(is.numeric)))替代硬编码的.[1:5]，适配列取值个数变化的场景，兼容性更强

内容的提问来源于stack exchange，提问作者ArturR2