如何在datasummary中添加列数与原表不一致的跨列行？

向datasummary生成的表格添加行时，若新增行的列数少于原表格，会触发如下错误：

Error in factory(dse, align = align, fmt = fmt, hrule = NULL, notes = notes, :
Assertion on 'add_rows' failed: Must have exactly 5 cols, but has 2 cols.

可正常运行的代码示例 #

library(modelsummary)
set.seed(123) # 保证结果可复现
data <- data.frame(
  x1 = rnorm(100),
  x2 = rnorm(100),
  y = rnorm(100)
)
summary_table <- datasummary(
  x1 + x2 + y ~ Mean + SD + Min + Max,
  data = data
)

new_row <- data.frame(
  First_variable = "New Row",
  Mean = 1,
  SD = 2,
  Min = 3,
  Max = 4
)

datasummary(
  x1 + x2 + y ~ Mean + SD + Min + Max,
  data = data, add_rows = new_row
)

无法运行的代码示例 #

new_row2 <- data.frame(
  First_variable = "New Row",
  Mean = "a bit of a longer text"
)

datasummary(
  x1 + x2 + y ~ Mean + SD + Min + Max,
  data = data, add_rows = new_row2
)

解决方法 #

要实现跨多列的行，需保证新增行的列数与原表格完全一致，空缺列用NA占位，再通过span参数指定跨列范围：

library(modelsummary)
set.seed(123)
data <- data.frame(
  x1 = rnorm(100),
  x2 = rnorm(100),
  y = rnorm(100)
)

# 构造列数匹配的新增行，空缺列填充NA
new_row_span <- data.frame(
  First_variable = "跨列说明文本",
  Mean = NA,
  SD = NA,
  Min = NA,
  Max = NA
)

# 使用span参数指定第2到第5列合并（对应Mean至Max列）
datasummary(
  x1 + x2 + y ~ Mean + SD + Min + Max,
  data = data,
  add_rows = new_row_span,
  span = list(c(2, 5))
)

如果需要调整跨列位置，只需修改span中的列索引范围即可。若想隐藏NA的显示，可搭配fmt参数将NA格式化为空字符串：

datasummary(
  x1 + x2 + y ~ Mean + SD + Min + Max,
  data = data,
  add_rows = new_row_span,
  span = list(c(2, 5)),
  fmt = function(x) ifelse(is.na(x), "", x)
)

内容的提问来源于stack exchange，提问作者Karl A