如何在reactablefmtr中动态生成color_tiles列定义？

你遇到的问题是R中闭包延迟绑定导致的——在for循环里创建的列定义，会共享循环变量i的最终值（也就是2），所以所有color_tiles的color_by都会指向最后一个class2列。

下面提供两种可行的解决方法：

方法1：用purrr::map代替for循环 #

map会为每个迭代创建独立的环境，避免变量共享问题：

library(reactable)
library(reactablefmtr)
library(dplyr)
library(purrr)

tile_column <- function(ref_column, name, show_text = FALSE) {
  # 不需要传入x，color_tiles会自动处理对应列的单元格数据
  colDef(
    show = TRUE,
    name = name,
    cell = color_tiles(
      color_by = ref_column,
      colors = c("red", "orange", "yellow", "limegreen", "green"),
      box_shadow = TRUE,
      show_text = show_text
    ),
    na = ""
  )
}

data <- tibble(
  name = paste("name", 1:10),
  val1 = paste("val1", 1:10),
  val2 = paste("val2", 1:10),
  class1 = sample(1:5, 10, replace = TRUE),
  class2 = sample(1:5, 10, replace = TRUE)
)

col_definitions1 <- list(
  name = colDef(show = TRUE, name = "Name")
)

# 用map生成列定义，每个迭代独立处理
col_definitions2 <- map(1:2, function(i) {
  tile_column(
    ref_column = paste0("class", i),
    name = paste("Column", i),
    show_text = TRUE
  )
}) %>% set_names(paste0("val", 1:2))

col_definitions <- c(col_definitions1, col_definitions2)

data %>%
  reactable(
    defaultColDef = colDef(show = FALSE),
    columns = col_definitions
  )

方法2：在for循环中强制求值变量 #

如果坚持用for循环，可以用force()函数强制当前迭代的ref_column值被立即求值，避免延迟绑定：

library(reactable)
library(reactablefmtr)
library(dplyr)
library(purrr)

tile_column <- function(ref_column, name, show_text = FALSE) {
  colDef(
    show = TRUE,
    name = name,
    cell = color_tiles(
      color_by = ref_column,
      colors = c("red", "orange", "yellow", "limegreen", "green"),
      box_shadow = TRUE,
      show_text = show_text
    ),
    na = ""
  )
}

data <- tibble(
  name = paste("name", 1:10),
  val1 = paste("val1", 1:10),
  val2 = paste("val2", 1:10),
  class1 = sample(1:5, 10, replace = TRUE),
  class2 = sample(1:5, 10, replace = TRUE)
)

col_definitions1 <- list(
  name = colDef(show = TRUE, name = "Name")
)

col_definitions2 <- list()
for (i in 1:2) {
  ref_col <- paste0("class", i)
  force(ref_col)  # 强制求值当前的ref_col
  col_definitions2 <- c(
    col_definitions2,
    list(tile_column(ref_column = ref_col, name = paste("Column", i), show_text = TRUE))
  )
}
col_definitions3 <- col_definitions2 |> set_names(paste0("val", 1:2))

col_definitions <- c(col_definitions1, col_definitions3)

data %>%
  reactable(
    defaultColDef = colDef(show = FALSE),
    columns = col_definitions
  )

另外注意：原tile_column函数里的x参数是多余的，color_tiles作为cell参数时，会自动接收对应列的单元格数据，不需要手动传入整个数据集。

内容的提问来源于stack exchange，提问作者eastclintw00d