r - 如何在R中为ggplot中的矩形自动查找序列的开始和结束

Question

我正在尝试用一些阴影矩形绘制一些数据。

数据框 df 如下所示:

df <- data.frame(time  = seq(0.1, 2, 0.1), 
                 speed = c(seq(0.5, 5, 0.5), seq(5, 0.5, -0.5)),
                 type  = c("a", "a", "b", "b", "b", "b", "c", "c", "c", "b", "b", "b", "b", "b", "c", "a", "b", "c", "b", "b"))

对于图中的矩形，我使用变量 xmin 和 xmax 定义了一个名为 dfRect 的对象。

dfRect <- data.frame(xmin = c(0.3, 1.0, 1.9), xmax = c(0.7, 1.5, 2.0))

问题是我必须为矩形的开始和结束手动找到 xmin 和 xmax。一个矩形在 type 列中 b 的时间序列的开始处开始 (xmin)，并在同一时间序列的结束处结束b。单个 b 可以忽略。

这是情节，因此您可以了解我要完成的工作:

ggplot() +
  geom_rect(data = dfRect, 
            aes(xmin = xmin, xmax = xmax, ymin = -Inf, ymax = Inf), 
            fill = "yellow", alpha = 0.4) +
  geom_line(data = df, aes(x = time, y = speed, color = type, group = 1), size = 3)

所以最后的问题是。如何自动化定义 xmin 和 xmax 的过程并自动创建 dfRect，这样我就不必自己定义它了？

Answer 1

这是一种使用游程编码的方法。

library(ggplot2)

df <- data.frame(time  = seq(0.1, 2, 0.1), 
                 speed = c(seq(0.5, 5, 0.5), seq(5, 0.5, -0.5)),
                 type  = c("a", "a", "b", "b", "b", "b", "c", "c", "c", "b", "b", "b", "b", "b", "c", "a", "b", "c", "b", "b"))

# Convert to runlength encoding
rle <- rle(df$type == "b")

# Ignoring the single "b"s
rle$values[rle$lengths == 1 & rle$values] <- FALSE

# Determine starts and ends
starts <- {ends <- cumsum(rle$lengths)} - rle$lengths + 1

# Build a data.frame from the rle
dfrect <- data.frame(
  xmin = df$time[starts],
  # We have to +1 the ends, because the linepieces end at the next datapoint
  # Though we should not index out-of-bounds, so we need to cap at the last end
  xmax = df$time[pmin(ends + 1, max(ends))],
  fill = rle$values
)

这个图显示了我们在上面的代码中所做的事情:

ggplot() +
  geom_rect(data = dfrect, 
            aes(xmin = xmin, xmax = xmax, ymin = -Inf, ymax = Inf, fill = fill), 
            alpha = 0.4) +
  geom_line(data = df, aes(x = time, y = speed, color = type, group = 1), size = 3)

要得到你想要的，你需要过滤掉 FALSEs。

ggplot() +
  geom_rect(data = dfrect[dfrect$fill,], 
            aes(xmin = xmin, xmax = xmax, ymin = -Inf, ymax = Inf), 
            alpha = 0.4, fill = "yellow") +
  geom_line(data = df, aes(x = time, y = speed, color = type, group = 1), size = 3)

如果您正在寻找可以为您计算的统计数据，请查看 here .免责声明:我编写了这个函数，它的作用与我上面发布的代码类似。