如何调整crosstalk复选框选项的排序规则？

要自定义crosstalk复选框的选项顺序，核心是把目标列转换成因子（factor），利用因子的水平顺序覆盖默认的字母排序逻辑，针对两种需求分别处理：

1. 按条目数量排序（水果类型） #

先统计每种水果的条目数，再将fruit_label转换为因子，水平顺序按条目数从多到少排列：

# 在数据处理环节添加因子转换
fruit_data <- data.frame(fruit = c(rep("apple", 4), rep("banana", 14), rep("pear", 12)),
                         quality = c(rep("awful", 10), rep("okay", 15), rep("amazing", 5))) |>
  group_by(fruit) |>
  mutate(fruit_label = paste0(fruit, " (", n(), ")")) |>
  # 提取括号内的数字，按降序排序后设置因子水平
  mutate(fruit_label = factor(fruit_label, 
                              levels = fruit_label |> unique() |> 
                                arrange(desc(as.numeric(sub(".*\\((\\d+)\\)", "\\1", .)))) |> 
                                pull())) |>
  ungroup()

这样水果类型的复选框就会按「香蕉(14) > 梨(12) > 苹果(4)」的顺序展示。

2. 按自定义因子水平排序（质量） #

直接将quality列转换为因子，指定自定义的水平顺序：

# 追加质量列的因子转换
fruit_data <- fruit_data |>
  mutate(quality = factor(quality, levels = c("amazing", "okay", "awful")))

质量复选框就会严格按照amazing > okay > awful的预设顺序排列。

完整可运行代码 #

---
title: "Ordering checkboxes from crosstalk"
---

```{r, echo=FALSE, message=FALSE}
library(dplyr)
library(crosstalk)
library(reactable)

fruit_data <- data.frame(fruit = c(rep("apple", 4), rep("banana", 14), rep("pear", 12)),
                         quality = c(rep("awful", 10), rep("okay", 15), rep("amazing", 5))) |>
  group_by(fruit) |>
  mutate(fruit_label = paste0(fruit, " (", n(), ")")) |>
  # 按条目数降序设置fruit_label因子水平
  mutate(fruit_label = factor(fruit_label, 
                              levels = fruit_label |> unique() |> 
                                arrange(desc(as.numeric(sub(".*\\((\\d+)\\)", "\\1", .)))) |> 
                                pull())) |>
  ungroup() |>
  # 自定义quality因子水平顺序
  mutate(quality = factor(quality, levels = c("amazing", "okay", "awful")))

shared_data <- SharedData$new(fruit_data)

# 注意：两个filter_checkbox的inputId要唯一，原代码重复会导致冲突
filter_checkbox("fruit_filter", "Type", shared_data, ~fruit_label)
filter_checkbox("quality_filter", "Quality", shared_data, ~quality)

reactable(shared_data,
          columns = list(
            fruit_label = colDef(show = FALSE)
          ))

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