条形图中scale_fill_manual和scale_x_discrete之间的冲突

Question

我试图控制绘制因子的顺序，并为它们指定特定的颜色。我对ggplot2相当新，是最近一次不情愿的基础图形转换。

以下是一个例子：

df <- data.frame(fac = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"),
                 val = c(10, 4, 7))  
ggplot(data = df, aes(x = fac, y = val, fill = fac)) +
       geom_bar(stat="identity", width=0.9) +
       scale_fill_manual(values=c("blue", "red", "orange")) +
       scale_x_discrete(limits = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"))

我想要＆＃34;服从Cowgod＆＃34;变成蓝色，＆＃34;三只小猪＆＃34;变成红色，和＃34;谷物杀手＆＃34;变成橙色怎么办呢？

Answer 1

您应该将列级别的顺序设置为一个因子。

df$fac <- factor(df$fac, levels = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"))  

现在颜色将分配给正确的名称。

（编辑注释：正如Gregor在评论中指出的那样，不再需要scale_x_discrete）

完整代码：

df <- data.frame(fac = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"),
                 val = c(10, 4, 7))

df$fac <- factor(df$fac, levels = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"))

ggplot(data = df, aes(x = fac, y = val, fill = fac)) +
  geom_bar(stat="identity", width=0.9) +
  scale_fill_manual(values=c("blue", "red", "orange"))

输出：

Answer 2

两种方式：

一个 - 按原样保留数据并使用命名向量进行填充values：

ggplot(data = df, aes(x = fac, y = val, fill = fac)) +
       geom_bar(stat="identity", width=0.9) +
       scale_fill_manual(values=c("Obey the Cowgod" = "blue", "Three Little Pigs" = "red", "Cereal Killer" = "orange")) +
       scale_x_discrete(limits = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"))

两个 - 按照您想要绘制的顺序按要素级别的顺序编辑数据，如Cihan的答案。我更喜欢这个，它是在我写的时候发布的：）

Answer 3

不要依赖于字符串的自动转换。在提供的代码中，级别以alpha顺序给出。

df <- data.frame(fac = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"),
                 val = c(10, 4, 7))

levels(df$fac)
# [1] "Cereal Killer"     "Obey the Cowgod"   "Three Little Pigs"

通过明确地将df$fac设置为您希望ggplot代码无需更改的级别的因素。

df <- data.frame(fac = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"),
                 val = c(10, 4, 7),
                 stringsAsFactors = FALSE)


df$fac <- factor(df$fac, levels = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"))

levels(df$fac)
# [1] "Obey the Cowgod"   "Three Little Pigs" "Cereal Killer"    

无需更改ggplot代码

library(ggplot2)
ggplot(data = df, aes(x = fac, y = val, fill = fac)) +
       geom_bar(stat="identity", width=0.9) +
       scale_fill_manual(values=c("blue", "red", "orange")) +
       scale_x_discrete(limits = c("Obey the Cowgod", "Three Little Pigs", "Cereal Killer"))