R语言如何转换宽表为长表并计算ymax/ymin适配环形图绘制需求

你可以直接使用tidyr包的pivot_longer函数完成宽表转长表，再分组计算ymin和ymax即可，完整可运行代码如下：

# 加载所需包
library(dplyr)
library(tidyr)

# 你原有的数据生成逻辑
position <- c("A", "A", "B", "C", "D", "E", "F", "F")
level <- c("P", "S", "J", "J", "P", "S", "S", "S")
car <- c("yes", "no", "no", "no", "no", "yes", "no", "yes")
car.data <- data.frame(position, level, car)

percent_car <- car.data %>%
  group_by(position) %>%
  summarise (car_yes = sum(car == "yes"), car_no = sum(car == "no")) %>%
  mutate(yes_perc = car_yes/(car_yes + car_no), no_perc = car_no/(car_yes + car_no))

# 转换为目标格式数据集
target_df <- percent_car %>%
  # 宽表转长表，按下划线拆分列名，拆分出yes/no标识和数值类型
  pivot_longer(
    cols = -position,
    names_to = c(".value", "yes_no"),
    names_sep = "_"
  ) %>%
  # 重命名为你需要的列名
  rename(
    Position = position,
    car_yes_no = car,
    yes_no_perc = perc
  ) %>%
  # 按职位分组计算环形图所需的上下边界值
  group_by(Position) %>%
  mutate(
    ymin = lag(cumsum(yes_no_perc), default = 0),
    ymax = cumsum(yes_no_perc)
  ) %>%
  ungroup()

# 查看输出结果
print(target_df, n = 12)

补充说明 #

如果需要调整yes和no的排列顺序，只需要在计算ymin/ymax前，将yes_no列转为自定义顺序的因子即可，示例：

# 调整为no在前，yes在后的顺序
target_df <- target_df %>%
  mutate(yes_no = factor(yes_no, levels = c("no", "yes"))) %>%
  arrange(Position, yes_no) %>%
  group_by(Position) %>%
  mutate(
    ymin = lag(cumsum(yes_no_perc), default = 0),
    ymax = cumsum(yes_no_perc)
  ) %>%
  ungroup()

内容的提问来源于stack exchange，提问作者Besz15