能否在gtsummary中多次调用add_n函数？

问题原因 #

连续调用add_n()时，第一次调用会修改表格结构，默认依赖的初始n列（总样本数列）被替换或移除，导致第二次调用add_n()找不到目标列，触发! Column "n" not found and was ignored.警告，且无法生成缺失值计数列。

解决方案 #

方案1：用add_stat()替代第二次add_n() #

add_stat()可直接针对每个变量计算缺失值数量，避免和add_n()的列冲突：

library(gtsummary)
library(dplyr)

iris %>%
  tbl_summary(type = list(where(is.logical) ~ "categorical"),
              missing_text = 'Missing',
              missing = 'ifany') %>%
  add_n(statistic = "{N_nonmiss}", col_label = "**N responses**") %>%
  add_stat(
    fns = ~n_miss(.),
    col_label = "**Missing**"
  )

方案2：通过modify_table_body一次性添加两列 #

直接修改表格主体数据，手动计算非缺失和缺失值数量，灵活性更高：

library(gtsummary)
library(dplyr)

iris %>%
  tbl_summary(type = list(where(is.logical) ~ "categorical"),
              missing_text = 'Missing',
              missing = 'ifany') %>%
  modify_table_body(
    ~.x %>%
      mutate(
        `N responses` = purrr::map_dbl(variable, ~sum(!is.na(iris[[.x]]))),
        Missing = purrr::map_dbl(variable, ~sum(is.na(iris[[.x]])))
      )
  ) %>%
  modify_header(`N responses` = "**N responses**", Missing = "**Missing**")

内容的提问来源于stack exchange，提问作者Julio