Plotly结合Crosstalk实现仅部分组显示标记的问题求助

核心是通过plotly的条件属性映射，在同一个trace中根据分组类型动态控制标记的显示，同时保留与crosstalk控件的关联。具体实现逻辑如下：

• 用crosstalk::SharedData包装数据集，确保交互控件能持续作用于图表
• 通过~ifelse()或dplyr::case_when()实现视觉属性的条件映射：
  • 标记：仅男女组显示标记，其余组设置为无标记（用"none"或尺寸0）
  • 线型：全体人口用实线，年龄组用不同虚线，男女组用实线

示例代码 #

library(plotly)
library(crosstalk)
library(dplyr)

# 构造示例人口统计数据
set.seed(123)
groups <- c("All", "18-30", "31-50", "51+", "男", "女")
dates <- seq.Date(as.Date("2020-01-01"), as.Date("2023-12-31"), by = "quarter")

df <- expand.grid(group = groups, date = dates) %>%
  mutate(population = case_when(
    group == "All" ~ rnorm(n(), 50000, 2000),
    group == "18-30" ~ rnorm(n(), 15000, 800),
    group == "31-50" ~ rnorm(n(), 20000, 1000),
    group == "51+" ~ rnorm(n(), 15000, 900),
    group == "男" ~ rnorm(n(), 25000, 1200),
    group == "女" ~ rnorm(n(), 25000, 1100)
  ))

# 创建SharedData用于crosstalk交互
sd <- SharedData$new(df, key = ~group, group = "population-controls")

# 组合交互控件与图表
bscols(
  list(
    filter_select("group", "选择显示组", sd, ~group, multiple = TRUE, selected = groups),
    filter_slider("date", "选择时间段", sd, ~date, step = 3)
  ),
  plot_ly(sd, x = ~date, y = ~population, color = ~group) %>%
    add_lines(
      # 条件控制标记：仅男女组显示圆形标记，其余组隐藏
      marker = list(
        symbol = ~ifelse(group %in% c("男", "女"), "circle", "none"),
        size = 6
      ),
      # 条件控制线型：全体用实线，年龄组用不同虚线，男女用实线
      linetype = ~case_when(
        group == "All" ~ "solid",
        group %in% c("18-30", "31-50", "51+") ~ c("dash", "dot", "dashdot")[match(group, c("18-30", "31-50", "51+"))],
        TRUE ~ "solid"
      ),
      text = ~group,
      hoverinfo = "text+x+y"
    ) %>%
    layout(title = "人口统计趋势（支持交互筛选）", xaxis = list(title = "时间"), yaxis = list(title = "人口数"))
)

关键细节说明 #

1. 标记隐藏：使用symbol = ~ifelse(...)指定"none"可完全隐藏标记，不会出现默认圆形；若该方式仍有警告，可替换为marker=list(size=~ifelse(group %in% c("男", "女"), 6, 0))，通过尺寸0实现隐藏
2. 线型区分：通过case_when()为不同分组匹配对应线型，确保年龄组之间有清晰区分
3. 交互保留：全程基于SharedData操作，crosstalk控件的筛选会直接同步到plotly图表，不会断开关联

内容的提问来源于stack exchange，提问作者EmilA