R语言dplyr实现数据框各列分类值计数与占比统计

注意事项 #

示例构造代码里的"NA"是字符串类型，不是R语言原生的缺失值NA，统计前需要先做转换，否则会把字符串"NA"识别为有效分类值，导致统计结果错误。

实现代码 #

# 加载依赖包
library(dplyr)
library(tidyr)
library(tibble)

# 构造示例数据，同时修正字符串"NA"为真实缺失值
id = seq(1:10)
A = c("NA","NA","cat","NA","cat","NA","NA","NA","cat","dog")
B = c("dog","NA","dog","NA","cat","dog","dog","dog","cat","cat")
C = c("NA","NA","cat","NA","NA","dog","NA","NA","cat","dog")
frame = tibble(id,A,B,C) %>%
  mutate(across(everything(), ~na_if(., "NA")))

# 核心统计逻辑
stat_result <- frame %>%
  # 将A/B/C三列从宽格式转为长格式，方便统一分组统计
  pivot_longer(cols = c(A, B, C), names_to = "col_name", values_to = "category") %>%
  # 剔除所有缺失值记录
  filter(!is.na(category)) %>%
  # 按列名、分类值分组，统计每个分类的出现次数
  count(col_name, category, name = "count") %>%
  # 按列分组，计算单列非缺失值总个数，拼接成要求的占比格式
  group_by(col_name) %>%
  mutate(
    total = sum(count),
    per = paste(count, total, sep = "/")
  ) %>%
  ungroup() %>%
  # 按列名拆分为独立的统计表
  group_split(col_name) %>%
  setNames(c("A", "B", "C"))

# 提取各列独立统计表，将分类值设为行名
table_A <- stat_result$A %>% select(-col_name, -total) %>% column_to_rownames("category")
table_B <- stat_result$B %>% select(-col_name, -total) %>% column_to_rownames("category")
table_C <- stat_result$C %>% select(-col_name, -total) %>% column_to_rownames("category")

输出结果 #

运行后三个统计表完全符合预期：

• table_A

• table_B

• table_C

说明 #

该实现没有硬编码cat/dog等具体分类值，只要是A/B/C列中出现的非缺失分类值都会被自动统计，后续新增分类值不需要修改代码逻辑。

内容的提问来源于stack exchange，提问作者Homer Jay Simpson