Secondary axis would not revert in ggplot version 3.1

[fa1therr]

I am trying to build a bar and line graphs on the same plot but with the different y-axis. The y1 axis has limits 0 to 100 and y2 axis has limits from 4 to 1 (reversed). Hence, I am adding secondary axis for the line plot through scale_y_continuous(limits=c(0,100), sec.axis=(...)). I build a linear mapping from 0-100 to 4-1 through lm() and then fit the coefficients to the formula of sec.axis(). This works fine in ggplot 2.2.1 version whatever limits for secondary axis I put. However, in ggplot 3.1 the limits would not revert. Is it a bug?

#Data generation
Dimension <- c("Sales", "Image", "PR","Presentations","Observation")

#ImportanceToPercentOfRespondents <- c(90%, 60%, 50%, 30%, 20%)
ImportanceToPercentOfRespondents <-c(0.9,0.6,0.5,0.3,0.2) * 100

# Was the Goal of each "Dimension" achieved?
AverageSuccess <- c(2.28, 1.82, 1.88,1.86, 2.77)
# 1 = yes, totally achived
# 5 = not at all
# Chart should however adapt to other scales as well.

df <- data.frame(Dimension, ImportanceToPercentOfRespondents, AverageSuccess)


# ----------- Font for chart -----------

# library(extrafont)
# font_import(pattern="Open Sans Semibold")
# font_import(pattern="Open Sans")


# ----------- Chart --------------------

require(ggplot2)    
# df <- data.frame(x = Dimension, y = abs(rnorm(5)*100)
df <- data.frame(x = Dimension, y = ImportanceToPercentOfRespondents)
df$y2 <- AverageSuccess #abs(rnorm(5))


#limits of the secondary axis - adjust boundaries when needed
y2_l = 1
y2_r = 4


#transformation factor
transf_fact <- 100/(y2_r - y2_l)



# Plot
p <- ggplot(data = df,aes(x = as.factor(x), y = y)) +
  geom_col(fill = '#A60032') +
  scale_x_discrete(limits = c("Sales","Image","PR","Presentations","Observation"))+
  
  # Shift y-axis by 1 unit, apply transformation factor and reflect symmetrically w.r.t. y = 50
  geom_line(aes(y = 100 -(df$y2 - (y2_l))*transf_fact, group = 1))


#We are looking for a linear mapping from 0-100 (left axis) to 4-1 (or 5-1, 6-1), adjust if necessary
m <- lm(y~x, data.frame(x=c(0,100), y = c(y2_r, y2_l)))

p + 
  # Add second OY axis; note the transformation back (division)
  scale_y_continuous(limits = c(0,100), 
                     sec.axis = sec_axis(trans = ~ (m$coefficients[[1]] + m$coefficients[[2]]*.), 
                                         name  = "Average Success (mean)"))+
  geom_label(aes(y = 100-(df$y2-y2_l)*transf_fact, 
                 label = round(y2, 2))) +
  xlab("Dimension") + 
  ylab("Important to test persons (%)")
# theme_bw() +
# theme(text=element_text(family = "Open Sans"))

[yutannihilation]

Is it a bug?

Yes.

This is a duplicate of #2978. I confirmed #3040 will fix this issue.

Dimension <- c("Sales", "Image", "PR","Presentations","Observation")
ImportanceToPercentOfRespondents <-c(0.9,0.6,0.5,0.3,0.2) * 100
AverageSuccess <- c(2.28, 1.82, 1.88,1.86, 2.77)
df <- data.frame(Dimension, ImportanceToPercentOfRespondents, AverageSuccess)

require(ggplot2)    
#> Loading required package: ggplot2
df <- data.frame(x = Dimension, y = ImportanceToPercentOfRespondents)
df$y2 <- AverageSuccess #abs(rnorm(5))

y2_l = 1
y2_r = 4

transf_fact <- 100/(y2_r - y2_l)

p <- ggplot(data = df,aes(x = as.factor(x), y = y)) +
  geom_col(fill = '#A60032') +
  scale_x_discrete(limits = c("Sales","Image","PR","Presentations","Observation"))+
  geom_line(aes(y = 100 -(df$y2 - (y2_l))*transf_fact, group = 1))

m <- lm(y~x, data.frame(x=c(0,100), y = c(y2_r, y2_l)))

p + 
  scale_y_continuous(limits = c(0,100), 
                     sec.axis = sec_axis(trans = ~ (m$coefficients[[1]] + m$coefficients[[2]]*.), 
                                         name  = "Average Success (mean)"))+
  geom_label(aes(y = 100-(df$y2-y2_l)*transf_fact, 
                 label = round(y2, 2))) +
  xlab("Dimension") + 
  ylab("Important to test persons (%)")

^{Created on 2019-02-25 by the reprex package (v0.2.1.9000)}

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/