使用ggplotly转换sf地图ggplot时出现sum(tickExists)类型错误

错误原因 #

这个错误是由于ggplotly解析使用geom_sf且完全隐藏坐标轴（比如theme_void()）的ggplot对象时，内部用于判断坐标轴刻度的变量tickExists被误赋值为列表类型（来自sf对象的geometry列），而sum()函数无法处理列表，因此触发报错。单纯将sf转为data.frame无法解决问题——要么丢失空间几何信息，要么geom_sf仍依赖sf属性，没触及ggplotly的内部解析逻辑。

解决办法 #

办法1：替换theme_void()为精细化的主题隐藏 #

不要用theme_void()直接移除所有坐标轴相关元素，而是手动隐藏可见元素但保留坐标轴的基础框架，避免ggplotly触发错误的刻度判断：

library(tigris)
library(ggplot2)
library(plotly)

# 获取并合并数据
states <- states(cb = TRUE)
custom_data <- data.frame(state = states$NAME, value = rnorm(nrow(states)))
states_merged <- merge(states, custom_data, by.x = "NAME", by.y = "state")

# 绘制地图，替换theme_void为精细化隐藏
p <- ggplot(states_merged) +
  geom_sf(aes(fill = value, text = paste("州:", NAME, "\n值:", value))) +
  theme(
    axis.text = element_blank(),
    axis.title = element_blank(),
    axis.ticks = element_blank(),
    panel.grid = element_blank(),
    panel.background = element_blank()
  )

# 转换为交互式图形
ggplotly(p, tooltip = "text")

办法2：将sf转为普通数据框，用geom_polygon绘图 #

把sf对象的几何坐标提取出来，用普通的geom_polygon替代geom_sf，让ggplotly能正常解析：

library(tigris)
library(ggplot2)
library(plotly)
library(tidyr)
library(purrr)

# 获取并合并数据
states <- states(cb = TRUE)
custom_data <- data.frame(state = states$NAME, value = rnorm(nrow(states)))
states_merged <- merge(states, custom_data, by.x = "NAME", by.y = "state")

# 转换sf为带坐标的普通数据框
states_df <- states_merged %>%
  mutate(coords = map(geometry, ~as.data.frame(.x))) %>%
  unnest(coords) %>%
  rename(x = X, y = Y)

# 用geom_polygon绘图
p <- ggplot(states_df) +
  geom_polygon(aes(x = x, y = y, group = NAME, fill = value, text = paste("州:", NAME, "\n值:", value))) +
  coord_fixed() +
  theme(
    axis.text = element_blank(),
    axis.title = element_blank(),
    axis.ticks = element_blank(),
    panel.grid = element_blank(),
    panel.background = element_blank()
  )

# 转换为交互式图形
ggplotly(p, tooltip = "text")

办法3：直接使用plotly原生空间绘图函数 #

绕过ggplotly转换，直接用plotly的add_sf函数处理sf对象：

library(tigris)
library(plotly)

# 获取并合并数据
states <- states(cb = TRUE)
custom_data <- data.frame(state = states$NAME, value = rnorm(nrow(states)))
states_merged <- merge(states, custom_data, by.x = "NAME", by.y = "state")

# 直接绘制交互式地图
plot_ly(states_merged, sf = TRUE, hoverinfo = "text", text = ~paste("州:", NAME, "\n值:", value)) %>%
  add_sf(color = ~value) %>%
  layout(
    showlegend = TRUE,
    xaxis = list(visible = FALSE),
    yaxis = list(visible = FALSE)
  )

内容的提问来源于stack exchange，提问作者Clark