在 plotly r 中独立于实际轨迹数据更改图例轨迹颜色

Question

如何独立于显示的内容更改图例中标记的颜色？

library(plotly)
library(data.table)
data=data.table(x=seq(1,100), y=rnorm(100), group = sample(letters[1:3], ,size=100, replace=T))
plot_ly(type="scatter", mode="markers") %>% 
  add_trace(data=data[group=="a"], x=~x,y=~y, name="Metric", legendgroup="m", showlegend=T, marker=list(size=10,color="red")) %>% 
  add_trace(data=data[group=="b"], x=~x,y=~y, name="Metric2", legendgroup="m", showlegend=F, marker=list(size=10,color="blue")) %>% 
  add_trace(data=data[group=="c"], x=~x,y=~y, name="Metric3", legendgroup="m", showlegend=F, marker=list(size=10,color="green")) %>% 
  layout(showlegend=T)

我希望图例中的单个红色圆圈是黑色的，而不是红色的，而我希望绘图上的点是 red/blue/green 当前的样子。

Answer 1

获得此效果的一种方法是将 add_markers() 包含在 visible = 'legendonly' 中，并为所有 add_trace() 行设置 showlegend = F：

library(plotly)
library(data.table)
data=data.table(x=seq(1,100), y=rnorm(100), group = sample(letters[1:3], ,size=100, replace=T))
plot_ly(type="scatter", mode="markers") %>% 
  add_markers(x = data$x[1], y = data$y[1], name = 'Metric', visible = 'legendonly', legendgroup='m', showlegend = T, marker = list(size = 10, color = 'black')) %>% 
  add_trace(data=data[group=="a"], x=~x,y=~y, name="Metric", legendgroup="m", showlegend=F, marker=list(size=10,color="red")) %>% 
  add_trace(data=data[group=="b"], x=~x,y=~y, name="Metric2", legendgroup="m", showlegend=F, marker=list(size=10,color="blue")) %>% 
  add_trace(data=data[group=="c"], x=~x,y=~y, name="Metric3", legendgroup="m", showlegend=F, marker=list(size=10,color="green")) %>%
  layout(showlegend=T)

这给

在 plotly r 中独立于实际轨迹数据更改图例轨迹颜色

Change legend trace color independently of actual trace data in plotly r

r

r-plotly