r - How to control ordering of stacked bar chart using identity on ggplot2

Question

Using this dummy data.frame

ts <- data.frame(x=1:3, y=c("blue", "white", "white"), z=c("one", "one", "two"))

I try and plot with category "blue" on top.

ggplot(ts, aes(z, x, fill=factor(y, levels=c("blue","white" )))) + geom_bar(stat = "identity")

gives me "white" on top. and

ggplot(ts, aes(z, x, fill=factor(y, levels=c("white", "blue")))) + geom_bar(stat = "identity")

reverses the colors, but still gives me "white" on top. How can I get "blue" on top?

Answer 1

For what it is worth, in ggplot2 version 2.2.1 the order of the stack is no longer determined by the row order in the data.frame. Instead, it matches the order of the legend as determined by the order of levels in the factor.

d <- data.frame(
  y=c(0.1, 0.2, 0.7),
  cat = factor(c('No', 'Yes', 'NA'), levels = c('NA', 'Yes', 'No')))

# Original order
p1 <- ggplot(d, aes(x=1, y=y, fill=cat)) +
  geom_bar(stat='identity')

# Change order of rows
p2 <- ggplot(d[c(2, 3, 1), ], aes(x=1, y=y, fill=cat)) +
  geom_bar(stat='identity')

# Change order of levels
d$cat2 <- relevel(d$cat, 'Yes')
p3 <- ggplot(d, aes(x=1, y=y, fill=cat2)) +
  geom_bar(stat='identity') 

grid.arrange(p1, p2, p3, ncol=3)

It results in the below plot: