`geom_area()` provides unclear error message when discrete x-axis is used

[83221n4ndr34]

I encountered an issue when using geom_area() in ggplot2: when the x-axis is discrete, the code either produces an irrelevant warning or fails to plot anything without an informative error message. Here are the details:

Discrete X-Axis with Specified Colors

When using geom_area() with discrete values for x-axis and specifying fill colors using scale_fill_manual(), the code produces the following warning:

Warning message:
No shared levels found between names(values) of the manual scale and the data's fill values.

Discrete X-Axis without specified colors

When using geom_area() discrete values for x-axis without specifying fill colors, ggplot2 silently fails to plot anything. There is no warning or error message to indicate the issue.

Suggestion

It would be helpful if the library could provide a more informative error message in these scenarios. For example, a message indicating that geom_area() does not support discrete x-axes would be very useful for diagnosing the issue.

Example code

Below is the demonstration of the problem with the example code.

library(ggplot2)

# example data
df <- data.frame(
  time = rep(1:10, each = 4),
  category = rep(c("A", "B", "C", "D"), times = 10),
  value = c(runif(10, 0, 5), runif(10, 0, 5), runif(10, 0, 5), runif(10, 0, 5))
)

colors <- c("A" = "#ff0000", "B" = "#ff90c7", "C" = "#167b2b", "D" = "#0000ff")

Plot with Numeric x-Axis (for control)

# time is numeric
p_numeric <- ggplot(df, aes(x = time, y = value, fill = category)) +
  geom_area(position = "stack") +
  scale_fill_manual(values = colors) +
  theme_minimal()

print(p_numeric)

Plot with Character x-Axis

# Convert time to character
df$time_chr <- as.character(df$time)

p_character <- ggplot(df, aes(x = time_chr, y = value, fill = category)) +
  geom_area(position = "stack") +
  scale_fill_manual(values = colors) +
  theme_minimal()

print(p_character)

Warning message: No shared levels found between names(values) of the manual scale and the data's fill values.

# Plot without specified colors
p_character_bis <- ggplot(df, aes(x = time_chr, y = value, fill = category)) +
  geom_area(position = "stack") +
  theme_minimal()

print(p_character_bis)

Same results with Factor x-Axis

# Convert time to factor
df$time_factor <- as.factor(df$time)

p_factor <- ggplot(df, aes(x = time_factor, y = value, fill = category)) +
  geom_area(position = "stack") +
  scale_fill_manual(values = colors) +
  theme_minimal()

print(p_factor)

# Plot without specified colors
p_factor_bis <- ggplot(df, aes(x = time_chr, y = value, fill = category)) +
  geom_area(position = "stack") +
  theme_minimal()

print(p_factor_bis)

Bonus: Forcing Discrete x-Axis to the Right Plot

p_numeric <- p_numeric + scale_x_discrete(name = "time_numeric_discrete")

print(p_numeric)

This plot renders correctly but without x-axis labels.

[teunbrand]

This is case 2 of the details section of ?aes_group_order. Briefly, your discrete x-axis counts towards grouping. If you manually set the groups correctly, all these issues disappear.

library(ggplot2)

df <- data.frame(
  time = rep(1:10, each = 4),
  category = rep(c("A", "B", "C", "D"), times = 10),
  value = c(runif(10, 0, 5), runif(10, 0, 5), runif(10, 0, 5), runif(10, 0, 5))
)

colors <- c("A" = "#ff0000", "B" = "#ff90c7", "C" = "#167b2b", "D" = "#0000ff")

ggplot(df, aes(x = as.character(time), y = value, fill = category, group = category)) +
  geom_area(position = "stack") +
  scale_fill_manual(values = colors) +
  theme_minimal()

^{Created on 2024-05-29 with reprex v2.1.0}

The error message is correct: there wasn't any fill column left after the incorrect grouping.