如何消除ggplot中的NA刻度并取消无数据月份的误差线绘制？

问题1：移除3月的误差线 #

geom_errorbar默认会遍历数据中所有行，包括3月的无数据行。只需在geom_errorbar中通过data参数筛选掉3月的数据，就能只给有数据的月份绘制误差线：

geom_errorbar(data = subset(stn1, Month != "MAR"), 
              aes(ymin=Temperature-Temperature_sd, ymax=Temperature+Temperature_sd), 
              width=.2, position=position_dodge(.9))

问题2：修复X轴第二个9月的NA刻度 #

你设置的factor level里的"SEPTEMBER"在数据集中没有对应的匹配值，或者数据中次年9月的Month字段值不是这个字符串，导致ggplot无法生成有效刻度标签。提供两种解决方式：

方式一：统一两年9月的标签 #

如果不需要区分两年的9月，先给数据中的次年9月标记一个临时标识，再通过scale_x_discrete设置统一的刻度标签：

# 预处理数据：给重复的SEP（次年9月）加临时后缀
stn1$Month <- ifelse(stn1$Month == "SEP" & duplicated(stn1$Month), "SEP_next", stn1$Month)

# 绘图时指定factor层级和刻度标签
bay <- ggplot(stn1, aes(x = factor(Month, level = c("SEP", "OCT", "NOV", "DEC", "JAN", "FEB",
                                                    "MAR", "APR", "MAY", "JUN", "JUL", "AUG", "SEP_next")), 
                        y = Temperature))+
  geom_bar(stat = "identity", position = position_dodge())+
  theme_classic()+
  xlab("Month")+
  ylab("Temperature°C")+
  geom_errorbar(data = subset(stn1, Month != "MAR"), 
                aes(ymin=Temperature-Temperature_sd, ymax=Temperature+Temperature_sd), 
                width=.2, position=position_dodge(.9))+
  scale_x_discrete(labels = c("SEP", "OCT", "NOV", "DEC", "JAN", "FEB",
                              "MAR", "APR", "MAY", "JUN", "JUL", "AUG", "SEP"))

方式二：区分两年9月的标签 #

如果需要明确区分两年的9月，直接通过scale_x_discrete给对应的层级设置自定义标签：

bay <- ggplot(stn1, aes(x = factor(Month, level = c("SEP", "OCT", "NOV", "DEC", "JAN", "FEB",
                                                    "MAR", "APR", "MAY", "JUN", "JUL", "AUG", "SEPTEMBER")), 
                        y = Temperature))+
  geom_bar(stat = "identity", position = position_dodge())+
  theme_classic()+
  xlab("Month")+
  ylab("Temperature°C")+
  geom_errorbar(data = subset(stn1, Month != "MAR"), 
                aes(ymin=Temperature-Temperature_sd, ymax=Temperature+Temperature_sd), 
                width=.2, position=position_dodge(.9))+
  scale_x_discrete(labels = c("SEP (Yr 1)", "OCT", "NOV", "DEC", "JAN", "FEB",
                              "MAR", "APR", "MAY", "JUN", "JUL", "AUG", "SEP (Yr 2)"))

完整可运行代码（以方式一为例） #

# 预处理数据
stn1$Month <- ifelse(stn1$Month == "SEP" & duplicated(stn1$Month), "SEP_next", stn1$Month)

# 绘制图表
bay <- ggplot(stn1, aes(x = factor(Month, level = c("SEP", "OCT", "NOV", "DEC", "JAN", "FEB",
                                                    "MAR", "APR", "MAY", "JUN", "JUL", "AUG", "SEP_next")), 
                        y = Temperature))+
  geom_bar(stat = "identity", position = position_dodge())+
  theme_classic()+
  xlab("Month")+
  ylab("Temperature°C")+
  geom_errorbar(data = subset(stn1, Month != "MAR"), 
                aes(ymin=Temperature-Temperature_sd, ymax=Temperature+Temperature_sd), 
                width=.2, position=position_dodge(.9))+
  scale_x_discrete(labels = c("SEP", "OCT", "NOV", "DEC", "JAN", "FEB",
                              "MAR", "APR", "MAY", "JUN", "JUL", "AUG", "SEP"))

print(bay)

内容的提问来源于stack exchange，提问作者Bailey Lin