使用ggplot2绘制堆叠条形图手动设色时遇报错及数据问题

Let's break down and fix both the error and warning you're seeing:

1. Fixing the "Continuous value supplied to discrete scale" Error #

The root issue here is that your row variable is a continuous integer (created with seq_len()), but scale_fill_manual expects a discrete categorical variable to map colors to. ggplot gets confused when you try to feed a continuous value into a discrete fill scale.

Solution: Convert row to a Factor #

When creating the row column, explicitly turn it into a factor instead of leaving it as an integer:

matrix$row <- as.factor(seq_len(nrow(matrix)))

This tells ggplot to treat each row as a distinct category, which plays nicely with your manual color scale.

2. Fixing the "Removed 83 rows containing missing values" Warning #

This warning pops up because your melted data (dat2) includes rows where value=0. When using position="fill" (which stacks bars proportionally), these 0-value rows contribute nothing to the bar height, so ggplot automatically removes them to avoid empty "slots" in the stack.

Solution (Optional but Recommended): Filter Out 0-Value Rows #

If you don't need to display 0s (which makes sense for proportional stacking), filter them out before plotting to eliminate the warning:

# Using dplyr (cleaner syntax)
dat2 <- melt(matrix, id.vars = "row") %>%
  dplyr::filter(value != 0)

# Or base R if you don't have dplyr installed
dat2 <- dat2[dat2$value != 0, ]

Full Fixed Code #

Putting it all together, here's the corrected code with a small typo fix in your color vector (missing hash for #00FFFF):

library(reshape2)
library(ggplot2)

# Create row as factor instead of integer
matrix$row <- as.factor(seq_len(nrow(matrix)))
# Melt data and filter out 0 values
dat2 <- melt(matrix, id.vars = "row") %>%
  dplyr::filter(value != 0)

# Your color vector (fixed hex color typo)
fill_2 <- c("yellow2", "hotpink4" , "brown", "rosybrown2", "seagreen", "royalblue" , "#FFFFCC","#CCFF66", "#99CC00", "#CC9900", "#FF9933", "#CC6666", "#FF3333", "black", "royalblue1", "royalblue2", "royalblue3", "royalblue4", "navy", "chocolate1", "chocolate2", "chocolate3" , "chocolate4" , "brown1" , "brown2", "brown3", "brown4", "azure", "azure1", "azure2", "azure3", "#00FFFF", "#33FFCC", "#66FF99", "#99FFFF", "#CCFFFF", "#006666" )

# Plot with corrected fill variable
ggplot(dat2, aes(y = value, x = variable, fill = row, group = row)) +
  geom_bar(stat = "identity", position = "fill") +
  guides(fill = FALSE) +
  scale_fill_manual(values = fill_2)

Quick Check: #

Since your fill_2 vector has exactly 37 colors (matching your 37-row dataset), converting row to a factor ensures each row gets the unique color you intended.

内容的提问来源于stack exchange，提问作者sdabral