将变量名称传递给制作图表的用户定义函数

Question

我在将变量名从共享数据集传递到我定义的函数时遇到问题。

我看过类似 R 问题的文档，并尝试实施一些解决方案但无济于事。

在下面的 reprex 中，我遇到错误“eval(expr, data, expr_env) 错误：找不到对象 'pct_increase'”

有人对此有解决方案吗？非常感谢。

代表：

library(tidyverse)
library(plotly)
library(crosstalk)

bar_tibble <- tibble(category = c("Cat 1", "Cat 1", "Cat 2", "Cat 2"),
                     pct_increase = c(0.17, 0.25, 0.64, 0.85),
                     week = c(1, 2, 1, 2)) 

bar_data_shared <- SharedData$new(bar_tibble, ~week)

SharedData$origData(bar_tibble)

make_bar_chart <- function(shared_data, xvar, yvar){
  
plot_ly(data = shared_data, hoverinfo = "none") %>%

  add_trace(
    x = ~xvar,
    y = ~yvar,
    type = "bar",
    transforms = list(
      list(
        type = "aggregate",
        groups = ~yvar,
        aggregations = list(
          list(
            target = "x", func = "avg", enabled = T)))))
}

make_bar_chart(bar_data_shared, pct_increase, category)

Answer 1

当你调用你的函数时，你必须使用真实的对象或字符串。换句话说，您必须将列名作为字符串传递。

从那里你运行进入另一个问题。一旦你创建了列名字符串，Plotly 将不会在没有额外干预的情况下将列连接到 SharedData 对象。

因为我不认为你想要参数名称作为轴的标签，所以我也添加了标签。

您可以执行以下操作来完成这项工作。

make_bar_chart <- function(shared_data, xvar, yvar){
  plot_ly(data = shared_data,
          hoverinfo = "none") %>%
    add_trace(
      x = ~.data[[xvar]],  # <- tell plotly 'connect to the data object!'
      y = ~.data[[yvar]],  # <- here, too
      type = "bar",
      transforms = list(
        list(
          type = "aggregate",
          groups = ~.data[[yvar]],  # <- here, too
          aggregations = list(
            list(
              target = "x", func = "avg", enabled = T))))
      ) %>% 
    layout(xaxis = list(title = xvar),
           yaxis = list(title = yvar))
}
# pass non-environment variables as strings
make_bar_chart(bar_data_shared, "pct_increase", "category")

将变量名称传递给制作图表的用户定义函数

Passing variable name to user-defined function that makes a plotly chart

r

plotly

crosstalk