Storing ggplots in loop makes all plots the last one

Question

I have been banging my head against the wall for hours trying to fix this. I am facing a similar issue to here but I am unable to work out my problem through the provided solution. Every time I run it, I can print() plot as I loop and see the correct ones but every loop, all plots in the list are replaced with the current one, until finally the list is just full of the last iteration. Can someone please help me work out why it is not working?

My reduced example (not working)

plotCount <- 5
multiplePlots <<- tibble(clockPlots = list(), index = double())
loopCount = 1
while(plotCount > 0) {
  
  plot <- ggplot()
  
  short <- tibble(val = c(1:12), x = 3 * sin(val * pi / 6), y = 3 * cos(val * pi / 6))
  long <- tibble(val = c(1:60), x = 4.5 * sin(val * pi / 30), y = 4.5 * cos(val * pi / 30))
  shorter <- short[loopCount,]
  longer <- long[loopCount,]
  plot <- plot + geom_segment(aes(x = 0, y = 0, xend = shorter[[2]], yend = shorter[[3]]), colour = "black", size = 3)
  plot <- plot + geom_segment(aes(x = 0, y = 0, xend = longer[[2]], yend = longer[[3]]), colour = "black", size = 2)
  
  print(plot)
  plotCount <- plotCount - 1
  loopCount <- loopCount + 1
  
  multiplePlots <<- multiplePlots %>%
    add_row(clockPlots = list(plot), index = loopCount)
}

multiplePlots[[1]][[1]]
multiplePlots[[1]][[2]]
multiplePlots[[1]][[3]]

After viewing solution in other document (not working either)

plotCount <- 5 
multiplePlots <<- tibble(plots = double(), index = double()) 
plot_list = list() 
loopCount = 1

short <- tibble(val = c(1:12), x = 3 * sin(val * pi / 6), y = 3 * cos(val * pi / 6)) 
long <- tibble(val = c(1:60), x = 4.5 * sin(val * pi / 30), y = 4.5 * cos(val * pi / 30))

while(plotCount > 0) {   
    shorter <- short[loopCount*2,]   
    longer <- long[loopCount*2,]   plot_list[[loopCount]] <- local({
        loopCount <- loopCount
        plot <- ggplot() + 
            geom_segment(aes(x = 0, y = 0, xend = shorter[[2]], yend = shorter[[3]]), colour = "black", size = 3) + 
            geom_segment(aes(x = 0, y = 0, xend = longer[[2]], yend = longer[[3]]), colour = "black", size = 2)   
    })
    plotCount <- plotCount - 1   
    loopCount <- loopCount + 1 
}

plot_list[[1]] 
plot_list[[2]] 
plot_list[[3]] 
plot_list[[4]] 
plot_list[[5]]

I have tried breaking it down so many times it barely even represents my original code but I can't work out what the issue is.

In the question you linked the accepted answer suggests to use a function and lapply. I had not problem to add your plots to a list that way. Bonus benefits are more readable/sensible code (the code above has multiple issues, e.g. why `<<-`, redundant variables, counters increased at weird points in the code, etc.). The function route is much much cleaner and less error prone... — dario, Oct 30 '21 at 07:47
@dario Could you please elaborate on why the counters are increased weirdly? I do not see why that is an issue? The reason for the <<- is because I pulled all this code out of a function and I wanted it to be global. This code is for a project where the goal is to only create one function so creating a function in a function would not suffice unfortunately — Zac Warham, Oct 30 '21 at 09:11
Assigning to the global env is almost always a code smell. The counter is increased before its used for the index. It's not wrong, just weird. At other places linebreaks or semicolons were missing (your last edit fixed that). — dario, Oct 30 '21 at 09:28

score 2 · Accepted Answer · answered Nov 03 '21 at 21:43

Here is the solution you are looking for

createPlot = function(i){
  dfs = tibble(x = 0, xend = 3*sin((1:12)[i]*pi/6), y=0, yend = 3*cos((1:12)[i]*pi/6))
  dfl = tibble(x = 0, xend = 4.5*sin((1:60)[i]*pi/30), y=0, yend =4.5*cos((1:60)[i]*pi/30))
  p1 = ggplot()+
    geom_segment(aes(x, y, xend = xend, yend = yend), data=dfs, colour = "black", size = 3)+
    geom_segment(aes(x, y, xend = xend, yend = yend), data=dfl, colour = "black", size = 3)
}

multiplePlots = tibble(index = 1:5) %>% 
  rowwise() %>% 
  mutate(clockPlots = list(createPlot(index)))

lapply(multiplePlots$clockPlots, plot)

Here are a few selected charts

Storing ggplots in loop makes all plots the last one

1 Answers1