ggplot2绘制柱状图时如何避免分类变量自动重排？

Great question! I totally get the frustration of having ggplot2 reorder your categorical variables alphabetically when you want to keep the original order from your dataframe. Let's break down two simpler alternatives to your current approach:

Option 1: Use scale_x_discrete() to Define Order Directly in ggplot #

This method lets you keep your dataframe as-is (no need to convert to a factor) by telling ggplot exactly what order to use for the X-axis right in the plot code.

Here's how to adjust your example:

library(ggplot2)

df <- data.frame(letter=c("ball","truck","apple","bear"), no= c(1,2,3,4))

ggplot(df, aes(letter, no)) + 
  geom_bar(stat="identity") +
  scale_x_discrete(limits = df$letter) # This forces the original order

The limits argument in scale_x_discrete() takes a vector of the category names in the exact order you want them displayed—here, we just pass the letter column directly from your dataframe to preserve its original sequence.

Option 2: Simplify Factor Conversion with forcats::fct_inorder() #

If you don't mind converting to a factor but want a cleaner, less error-prone way than manually setting levels, the forcats package (part of the tidyverse) has a handy function called fct_inorder(). It automatically sets the factor levels to the order the values first appear in the dataframe.

Example code:

library(ggplot2)
library(forcats) # Load the forcats package

df <- data.frame(letter=c("ball","truck","apple","bear"), no= c(1,2,3,4))

df$letter <- fct_inorder(df$letter) # No need to manually specify levels!

ggplot(df, aes(letter, no)) + 
  geom_bar(stat="identity")

This cuts down on the code you need to write compared to your original factor conversion, and it's less likely to have typos or ordering mistakes.

Quick Comparison #

• Option 1 is perfect if you want to avoid modifying your dataframe at all—all the ordering logic stays in the plot code.
• Option 2 is great if you plan to reuse the ordered categorical variable elsewhere in your analysis, since it stores the order directly in the dataframe.

内容的提问来源于stack exchange，提问作者user8369515