如何在ggplot中引用管道传递的对象设置scale_fill_gradient2中点？

The issue you're hitting is that scale_fill_gradient2's midpoint parameter expects a static numeric value, not an expression that references columns from your piped data frame directly. Here's how to fix it, plus an explanation of why your original attempts failed:

Working Solutions #

1. Precompute the Median (Simplest Approach) #

Calculate the median of your modified Sepal.Length first, then use that scalar value in the scale:

# Calculate the median of the scaled Sepal.Length
median_scaled_sepal <- median(iris$Sepal.Length * 100)

# Build your plot with the precomputed median
iris %>% 
  mutate(Sepal.Length = Sepal.Length * 100) %>% 
  ggplot(aes(x = Sepal.Length, y = Sepal.Width, fill = Sepal.Length)) +
  geom_col() +
  scale_fill_gradient2(
    low = "yellow", 
    high = "green", 
    mid = "red", 
    midpoint = median_scaled_sepal
  )

2. Use Magrittr's Exposition Pipe (%$%) #

If you want to keep everything in a single pipe chain, use the %$% pipe (included with dplyr by default) to expose your data frame's columns to the subsequent expression:

iris %>% 
  mutate(Sepal.Length = Sepal.Length * 100) %$%  # Exposes columns to the next line
  ggplot(aes(x = Sepal.Length, y = Sepal.Width, fill = Sepal.Length)) +
  geom_col() +
  scale_fill_gradient2(
    low = "yellow", 
    high = "green", 
    mid = "red", 
    midpoint = median(Sepal.Length)
  )

3. Extract the Median from the Modified Data Frame #

Modify the data first, then pull the median value before plotting:

# Modify the data and store it
modified_iris <- iris %>% mutate(Sepal.Length = Sepal.Length * 100)

# Extract the median
median_val <- modified_iris %>% pull(Sepal.Length) %>% median()

# Plot using the extracted median
modified_iris %>% 
  ggplot(aes(x = Sepal.Length, y = Sepal.Width, fill = Sepal.Length)) +
  geom_col() +
  scale_fill_gradient2(
    low = "yellow", 
    high = "green", 
    mid = "red", 
    midpoint = median_val
  )

Why Your Original Code Failed #

• When you tried median(.$Sepal.Length): The . inside scale_fill_gradient2 refers to the ggplot object, not the piped data frame. Scale functions don't inherit the data context from the pipe chain.
• When you tried median(Sepal.Length): Sepal.Length only exists inside your modified data frame (from the mutate step), not in the global environment where scale_fill_gradient2 tries to evaluate it.

内容的提问来源于stack exchange，提问作者J. Doe