Plotly R：通过按钮筛选柱状图并隐藏无关EPISD类别

问题核心是：仅用restyle控制trace显示/隐藏时，x轴会保留所有trace中的EPISD值，导致无关类别残留。需要通过update方法同时更新trace可见性和x轴的类别范围。

修改后的代码 #

library(plotly)
library(data.table)

# 数据
dt <- data.table(
  EPISD = c("A", "A", "B", "B", "C", "D", "D"),
  ENTRP = c(
    "company 1",
    "company 2",
    "company 1",
    "company 2",
    "company 2",
    "company 1",
    "company 2"
  ),
  prct_na = c(100, 0, 100, 5.3, 100, 94, 100)
)

# 预先提取各公司的EPISD类别
comp1_episd <- unique(dt[ENTRP == "company 1"]$EPISD)
comp2_episd <- unique(dt[ENTRP == "company 2"]$EPISD)

plot_ly() %>%
  add_trace(
    data = dt[ENTRP == "company 1"],
    x = ~EPISD,
    y = ~prct_na,
    name = 'company 1',
    type = 'bar',
    visible = TRUE
  ) %>%
  add_trace(
    data = dt[ENTRP == "company 2"],
    x = ~EPISD,
    y = ~prct_na,
    name = 'company 2',
    type = 'bar',
    visible = FALSE
  ) %>%
  layout(
    updatemenus = list(
      list(
        type = "buttons",
        direction = "down",
        buttons = list(
          list(
            method = "update",
            args = list(
              list(visible = c(TRUE, FALSE)),  # 显示company1，隐藏company2
              list(xaxis = list(
                categoryarray = comp1_episd,
                categoryorder = "array"
              ))
            ),
            label = "company 1"
          ),
          list(
            method = "update",
            args = list(
              list(visible = c(FALSE, TRUE)),  # 显示company2，隐藏company1
              list(xaxis = list(
                categoryarray = comp2_episd,
                categoryorder = "array"
              ))
            ),
            label = "company 2"
          )
        )
      )
    )
  )

关键改动说明 #

• 把按钮的method从restyle改为update：update支持同时修改trace属性和布局属性（比如x轴），而restyle仅能修改trace属性。
• 预先提取各公司对应的EPISD唯一值：保证切换时x轴只显示当前公司存在的类别。
• 在按钮的args中添加x轴更新参数：通过categoryarray指定x轴要展示的类别，categoryorder = "array"确保x轴按指定顺序排列类别。

这样切换按钮时，不仅会显示对应公司的柱状图，x轴也会自动过滤无关的EPISD类别，只保留当前公司的有效类别。

内容的提问来源于stack exchange，提问作者nimliug