升级至ggplot2 3.3.4版本后绘图出现意外颜色变化问题

问题原因 #

ggplot2 3.3.4版本调整了离散色阶的默认匹配逻辑：当scale_colour_manual()只指定breaks但未显式指定limits或给values命名时，未出现在breaks中的因子水平会无法正确匹配预设颜色，最终出现颜色错位、变灰等异常。

效果差异对比 #

解决方法 #

有两种兼容两个版本的修正方案，任选其一即可：

• 方案1：为颜色值命名，固定水平与颜色的对应关系
  把scale_colour_manual里的values转为命名向量，每个颜色对应固定的stadio水平，不受版本逻辑影响，适配性最强。
• 方案2：显式指定limits参数
  在scale_colour_manual中添加limits参数，指定所有要匹配的因子水平，保证11个颜色按顺序对应1-11的stadio值。

修正后的完整代码 #

library(RColorBrewer)
library(ggplot2)

df <- structure(list(name = structure(c(11L, 11L, 11L, 11L, 11L, 11L, 
11L, 11L, 11L, 11L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 12L, 
12L, 12L, 12L, 12L, 12L, 12L, 12L, 12L, 12L, 10L, 10L, 10L, 10L, 
10L, 10L, 10L, 10L, 10L, 10L), .Label = c("5303 B", "5303 A", 
"5302 B", "5302 A", "5301 B", "5301 A", "5202 B", "5202 A", "5201 B", 
"5201 A", "5101 B", "5101 A"), class = "factor"), stadio = c(2, 
2, 4, 4, 6, 6, 8, 8, 10, 10, 2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 
1, 1, 3, 3, 7, 7, 9, 9, 11, 11, 1, 1, 3, 3, 7, 7, 9, 9, 11, 11
), variable = structure(c(1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L), .Label = c("start_date", "end_date"), class = "factor"), value = c("15/10/2021", "03/01/2022", 
"03/01/2022", "30/03/2022", "30/03/2022", "30/03/2023", "30/03/2023", 
"15/09/2023", "15/09/2023", "29/12/2023", "15/10/2021", "03/01/2022", 
"03/01/2022", "30/03/2022", "30/03/2022", "30/03/2023", "30/03/2023", 
"15/09/2023", "15/09/2023", "29/12/2023", "26/11/2021", "14/01/2022", 
"14/01/2022", "30/06/2022", "30/06/2022", "30/03/2023", "30/03/2023", 
"15/09/2023", "15/09/2023", "29/12/2023", "26/11/2021", "14/01/2022", 
"14/01/2022", "30/06/2022", "30/06/2022", "30/03/2023", "30/03/2023", 
"15/09/2023", "15/09/2023", "29/12/2023"), rating = structure(c(1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("2", "3"), class = "factor")), row.names = c(NA, 
-40L), class = c("tbl_df", "tbl", "data.frame"))

ggplot(df, aes(as.POSIXct(as.Date(value, "%d/%m/%Y")), name, colour = factor(stadio,levels=1:11))) +
  guides(alpha = "none",
         colour = guide_legend(
           override.aes = list(size = 4))) +
  geom_line(aes(size=rating, alpha=rating)) +
  scale_alpha_manual(values=c(.5,1)) +
  labs(colour="(Actual/Baseline)", x = "", y = "") +
  # 方案1：给values命名，固定对应关系，二选一即可
  scale_colour_manual(breaks = c("1", "3", "7","9","11"), values = setNames(brewer.pal(11,"PiYG"), 1:11)) +
  # 方案2写法：显式指定limits，注释掉方案1后放开下方代码即可
  # scale_colour_manual(breaks = c("1", "3", "7","9","11"), values = brewer.pal(11,"PiYG"), limits = as.character(1:11)) +
  scale_size_manual(breaks = levels(df$rating),  values = as.integer(levels(df$rating)), guide = "none") 

内容的提问来源于stack exchange，提问作者Yorgos