将expss表格转为huxtable后出现空单元格问题求助

问题描述 #

制作报告时偏好expss的表格格式，但因huxtable能在R Markdown中实现表格样式自定义与自动添加标题，选择将expss表格转为huxtable格式。但转换后出现异常：原expss表格中Valid N列有数值，转为huxtable后该列单元格全部为空。

重现代码 #

expss表格生成代码 #

tab4 = m1 %>% 
       tab_cells(age, num_year, years) %>%
       tab_cols(sex) %>% 
       tab_stat_fun(Mean = w_mean, Median =w_median, "Std. dev." = w_sd, "Valid N" = w_n ,method = list) %>%
       tab_pivot() |>
       tab_caption("Summary statistics")
tab4

huxtable格式转换代码 #

as_hux(tab4) |>
   set_bottom_border(col = c(2,3,4,5,6,7,8,9), row = everywhere) |>
   set_bold(col = 1, row=everywhere) |>
   set_bold(col = c(2,3,4,5,6,7,8,9), row = c(1,2,3)) |>
   set_row_height(c(.2, .2, .2)) |>
   set_align(col = c(2,3,4,5,6,7,8,9), row = everywhere, value = "centre") 

测试数据集（前20行） #

structure(list(age = structure(c(32, 45, 53, 41, 68, 41, 42,
52, 52, 35, 37, 64, 54, 54, 38, 56, 33, 41, 34, 68), label = "Age (in years)", class = c("labelled",
"numeric")), years = structure(c(3, 16, 20, 17, 20, 12, 11, 15,
20, 6, 1, 32, 4, 7, 12, 33, 12, 12, 10, 40), label = "Experience in field (in years)", class = c("labelled",
"numeric")), num_year = structure(c(3, 20, 27, 17, 40, 12, 6,
9, 20, 8, 11, 32, 8, 20, 12, 33, 12, 4, 10, 40), label = "Experience with current office (in years)", class = c("labelled",
"numeric"))), row.names = c(NA, -20L), class = c("tbl_df", "tbl",
"data.frame"))

可能原因及解决建议 #

1. 特殊类型未被正确解析 #

expss生成的Valid N列可能是labelled类型，huxtable的as_hux()函数对该类型的数值解析存在问题，导致显示为空。

• 解决：转换前先将Valid N列转为普通数值：

tab4 = m1 %>% 
       tab_cells(age, num_year, years) %>%
       tab_cols(sex) %>% 
       tab_stat_fun(Mean = w_mean, Median =w_median, "Std. dev." = w_sd, "Valid N" = w_n ,method = list) %>%
       tab_pivot() |>
       mutate(across(contains("Valid N"), as.numeric)) |> # 新增转换步骤
       tab_caption("Summary statistics")

2. method = list导致结构不兼容 #

tab_stat_fun中使用method = list会改变统计结果的存储结构，huxtable转换时无法识别该结构下的数值。

• 解决：移除method = list参数，使用默认设置生成表格后再转换。

3. 格式设置代码干扰 #

先跳过所有格式设置，直接输出as_hux(tab4)，检查Valid N列是否有值，排除是格式设置中列索引错误导致的数值隐藏。

内容的提问来源于stack exchange，提问作者Nadeem