如何调整簇状geom_col的填充顺序以匹配图例顺序？

问题原因 #

你在scale_fill_brewer中设置的limits参数仅能调整图例的展示顺序，不会改变Season变量本身的因子水平，而簇状条形的排列优先级由变量的因子水平决定。Season默认会按照字母顺序排序（Autumn首字母A最靠前，会排在每组第一个），因此会出现条形顺序与图例顺序不匹配的问题。

解决方法 #

方案1（推荐）：预处理数据，提前定义因子顺序 #

在调用ggplot前，先将Season转换为指定顺序的有序因子，后续绘图会直接沿用该顺序：

# 指定Season的因子水平，顺序和你需要的图例/条形顺序完全匹配
data1day$Season <- factor(data1day$Season, levels = c("Summer", "Autumn", "Winter", "Spring"))

预处理完成后，简化原ggplot代码即可，无需额外调整scale参数，也不需要用arrange排序数据：

plot <-
  ggplot(data1day, aes(x = Set, y = 100 * Density, fill = Season)) +
  geom_col(position = 'dodge', color = 'black') +
  theme_classic() +
  labs(x = "Subset of sites surveyed", y = "Density (%)") +
  scale_x_discrete(limits = c("Full set", "Set 2", "Set 3", "Set 4")) +
  scale_fill_brewer(palette = "Set2") +
  theme(text = element_text(size = 14, family = "serif", colour = "black")) +
  theme(axis.text.x = element_text(colour = "black")) +
  theme(axis.text.y = element_text(colour = "black"))

方案2：不修改原数据，直接在绘图逻辑中指定顺序 #

如果你不想改动原数据集，可以直接在aes的fill参数中定义因子顺序，效果和方案1一致：

plot <-
  ggplot(data1day, aes(x = Set, y = 100 * Density, fill = factor(Season, levels = c("Summer", "Autumn", "Winter", "Spring")))) +
  geom_col(position = 'dodge', color = 'black') +
  theme_classic() +
  labs(x = "Subset of sites surveyed", y = "Density (%)", fill = "Season") +
  scale_x_discrete(limits = c("Full set", "Set 2", "Set 3", "Set 4")) +
  scale_fill_brewer(palette = "Set2") +
  theme(text = element_text(size = 14, family = "serif", colour = "black")) +
  theme(axis.text.x = element_text(colour = "black")) +
  theme(axis.text.y = element_text(colour = "black"))

两种方案运行后，每组的条形顺序都会和图例顺序完全对齐，颜色和数值的对应关系也不会发生变化。

内容的提问来源于stack exchange，提问作者Tormey R