自定义tidyverse函数中dplyr::summarise()报错排查求助

错误原因 #

调用函数时用vars()包裹变量参数，但dplyr的{{}}（curly curly）语法期望接收裸变量名或单个quosure，vars()返回的是quosure列表，导致ifelse无法识别参数类型，触发报错。

修复步骤 #

1. 修改函数调用方式 #

去掉vars()包裹，直接传入裸变量名：

# 修正后的调用
df.vetsite1.2021 %>% make_vet_site_table_row(var1 = var1_final_denom, var2 = var2_for_cnt)
df.vetsite2.2021 %>% make_vet_site_table_row(var1 = var1_beginning_denom, var2 = var2_for_cnt)

2. 优化函数内部逻辑（可选，更简洁健壮） #

替换基础ifelse为dplyr::if_else（类型检查更严格），也可以用逻辑值乘法简化计算：

make_vet_site_table_row <- function(df, var1, var2) {
  df %>% 
    dplyr::summarise(
      "Total number of vets"                     = n(),
      "Number of customers with at least 11 dogs" = sum(dplyr::if_else({{var1}} >= 11, {{var2}}, 0)),
      "Number of customers with at least 20 dogs" = sum(dplyr::if_else({{var1}} >= 20, {{var2}}, 0)),
      "Total number of dogs"                       = sum({{var1}}),
      "Mean number of patients (by vet)"         = round(mean({{var1}}), digits = 2),
      "Range of the number of patients (by vet)" = max({{var1}}) - min({{var1}}),
      "Smallest and largest number of dogs (by vet)" = paste0(min({{var1}}), ", ", max({{var1}}))
    )
}

3. 验证结果 #

用示例数据集测试：

# 构造示例数据集
df.vetqm <- tibble::tibble(
  site = c("site1", "site2", "site2", "site2", "site3", "site3"),
  var1_final_denom = c(3, 66, 23, 45, 542, 68),
  var2_for_cnt = c(3, 65, 45, 48, 423, 63),
  year = rep("2021", 6)
)

# 分组预处理
df.vetsite.2021 <- df.vetqm %>% filter(year=="2021" & var1_final_denom>0) %>% group_by(site)

# 调用生成表格
df.vetsite.2021 %>% make_vet_site_table_row(var1 = var1_final_denom, var2 = var2_for_cnt)

输出结果会自动保留site分组列，符合期望格式：

内容的提问来源于stack exchange，提问作者Tom