ggplot: Assign axis labels using variable pairs did not work

tidyverse
1

I wonder why the last case didn't work while the 1st and 2nd ones worked. I'm sure I'm missing something obvious here. Thank you for any help!


library(ggplot2)

p <- ggplot(mtcars, aes(wt, mpg))

lab1 <- as.character(seq(1, 5, 1))
lab2 <- c("One", "Two", "Three", "Four", "Five")

#1:works
p + geom_point(aes(size = qsec, fill = qsec), shape = 21, alpha = 0.7) +
  scale_fill_viridis_c(guide = "legend") +
  scale_size_continuous(range = c(1, 15)) +
  scale_x_continuous(breaks = seq(1, 5, 1),
                     labels = lab2)



#2:works too
p + geom_point(aes(size = qsec, fill = qsec), shape = 21, alpha = 0.7) +
  scale_fill_viridis_c(guide = "legend") +
  scale_size_continuous(range = c(1, 15)) +
  scale_x_continuous(breaks = seq(1, 5, 1),
                     labels = c("1" = "One", 
                                "2" = "Two",
                                "3" = "Three",
                                "4" = "Four",
                                "5" = "Five"))


#3:error
p + geom_point(aes(size = qsec, fill = qsec), shape = 21, alpha = 0.7) +
  scale_fill_viridis_c(guide = "legend") +
  scale_size_continuous(range = c(1, 15)) +
  scale_x_continuous(breaks = seq(1, 5, 1),
                     labels = c(lab1[1] = lab2[1], 
                                lab1[2] = lab2[2],
                                lab1[3] = lab2[3],
                                lab1[4] = lab2[4],
                                lab1[5] = lab2[5]))

Error: unexpected '=' in:
"  scale_x_continuous(breaks = seq(1, 5, 1),
                     labels = c(lab1[1] ="
>                                 lab1[2] = lab2[2],
Error: unexpected ',' in "                                lab1[2] = lab2[2],"
>                                 lab1[3] = lab2[3],
Error: unexpected ',' in "                                lab1[3] = lab2[3],"
>                                 lab1[4] = lab2[4],
Error: unexpected ',' in "                                lab1[4] = lab2[4],"
>                                 lab1[5] = lab2[5]))
Error: unexpected ')' in "                                lab1[5] = lab2[5])"
2

You don't need a named vector for labels. Your 2nd example works fine without the names in the vector.

p + geom_point(aes(size = qsec, fill = qsec), shape = 21, alpha = 0.7) +
  scale_fill_viridis_c(guide = "legend") +
  scale_size_continuous(range = c(1, 15)) +
  scale_x_continuous(breaks = seq(1, 5, 1),
                     labels = c("One", 
                                "Two",
                                "Three",
                                "Four",
                                "Five"))

The order of the breaks and labels is what creates the correct 'mapping' of new labels to the axis.

1 Like
3

What I wanted was to supply the exact label pairs. Maybe there were cases where labels could be mismatched because of the wrong orders? Thank you

4

Sure, you should use caution about the order, but using a named vector in labels doesn't give you any extra protection because the names are ignored in that named vector.

If you get rid of the breaks argument in #2, you get an error

p + geom_point(aes(size = qsec, fill = qsec), shape = 21, alpha = 0.7) +
  scale_fill_viridis_c(guide = "legend") +
  scale_size_continuous(range = c(1, 15)) +
  scale_x_continuous(labels = c("1" = "One", 
                                "2" = "Two",
                                "3" = "Three",
                                "4" = "Four",
                                "5" = "Five"))

Error in f(..., self = self) : Breaks and labels are different lengths
1 Like
5

I thought that ggplot2 would have option to prevent the mismatch between the actual labels and the axis values like this case

p + geom_point(aes(size = qsec, fill = qsec), shape = 21, alpha = 0.7) +
  scale_fill_viridis_c(guide = "legend") +
  scale_size_continuous(range = c(1, 15)) +
  scale_x_continuous(breaks = seq(1, 5, 1),
                     labels = c("Two", 
                                "One",
                                "Five",
                                "Four",
                                "Three"))

6

When you venture into the realm of custom labels, it is up to you to program defensively to prevent such an error.

2 Likes