在R中如何让table1按列总数值而非行数计算分组百分比

你遇到的问题核心是table1默认按行数统计占比，只需新增weights参数指定权重为count列，即可自动按count列的实际统计值计算总和与对应占比，默认输出格式即为「人数总和(占比%)」，无需额外自定义渲染函数，修改后代码如下：

set.seed(123)
dat <- data.frame(
  site = factor(sample(c("A", "B", "C", "D"), 100, replace = TRUE)),
  count = (sample(1:150, 100, replace = TRUE)),
  first.name = factor(sample(c("John", "Sue", "Bob", "Mary", "Cara"), 100, replace = TRUE)),
  last.name =  factor(sample(c("Williams", "Smith", "Lee"), 100, replace = TRUE)))

library(table1)
tab <- table1(~ last.name + first.name | site, 
              data = dat, 
              weights = count, # 核心参数：指定count为权重，按实际统计值计算
              digits= 1) 
tab

如果需要自定义数值格式、调整显示逻辑，也可以自行编写render.categorical函数实现。

如果需要更自由的表格格式调整，比如合并单元格、添加统计标识、自定义表头样式等，可以先手动汇总统计结果，再输出为规范统计表，代码如下：

library(dplyr)
library(flextable)

# 汇总姓氏分站点统计结果
last_res <- dat %>%
  group_by(site, last.name) %>%
  summarise(total = sum(count), .groups = "drop_last") %>%
  mutate(pct = round(100 * total / sum(total), 1),
         display = paste0(total, " (", pct, "%)")) %>%
  select(-total, -pct) %>%
  tidyr::pivot_wider(names_from = site, values_from = display) %>%
  rename("分类项" = last.name) %>%
  mutate(分类项 = paste0("姓氏：", 分类项))

# 汇总名字分站点统计结果
first_res <- dat %>%
  group_by(site, first.name) %>%
  summarise(total = sum(count), .groups = "drop_last") %>%
  mutate(pct = round(100 * total / sum(total), 1),
         display = paste0(total, " (", pct, "%)")) %>%
  select(-total, -pct) %>%
  tidyr::pivot_wider(names_from = site, values_from = display) %>%
  rename("分类项" = first.name) %>%
  mutate(分类项 = paste0("名字：", 分类项))

# 合并结果输出三线表格式的规范统计表
final_tab <- bind_rows(last_res, first_res) %>%
  flextable() %>%
  theme_booktabs() %>%
  autofit()

final_tab

内容的提问来源于stack exchange，提问作者SageandSun