Using `position_dodge()` to vertically dodge `geom_line()` on a factor y-axis

[DanChaltiel]

Hi,

I'm trying to connect vertically dodged points using geom_line(), but the lines don't follow the dodged positions correctly.

I'm using the orientation argument introduced in v4.0, but I think the problem is rather caused by position and geom_line() both using the same group aesthetic.

An example speaks louder than words, so here is the code to reproduce the behaviour.

library(tidyverse)
d = tibble::tibble(
  y_row = rep(c("d1", "d2"), each = 9L),
  source = rep(rep(c("s1", "s2", "s3"), 2), rep(4:2, 2)),
  value = c(1, 2, 3, 4, 2, 3, 4, 3, 4, 2, 3, 4, 5, 3, 4, 5, 4, 5),
)

pd = position_dodge(width=0.4, orientation="y")
# pd = position_dodge2(width=0.4)
ggplot(d) +
  aes(x=value, y=y_row, color=source) +
  geom_point(aes(group=source), position=pd) +
  geom_line(aes(group=y_row), position=pd)

^{Created on 2025-11-01 with reprex v2.1.1}

Expected output: points should be connected by a line of the same color

As v4.0 introduced auxiliary aesthetics nudge_x, nudge_y and order, maybe another could be added to dissociate the grouping for geom_line() and position_dodge()?

Thanks a lot for your time and for maintaining ggplot2!
Happy to help if you need more examples or testing.

Addendum

Using pd = position_dodge2(width=0.4) gives the same result.

Changing the group in geom_line() gives a very different plot, and the dodging is not correct either:

ggplot(d) +
  aes(x=value, y=y_row, color=source, group=y_row) +
  geom_point(aes(group=source), position=pd) +
  geom_line(aes(group=source))

The mere fact of adding/removing the position dodge with Y orientation changes the x-axis somehow:

ggplot(d) +
  aes(x=value, y=y_row, color=source, group=y_row) +
  geom_point(aes(group=source), position=pd) +
  geom_line(aes(group=source), position=pd)

[smouksassi]

see example of using orientation in the geom directly the position will inherit it.

library(ggplot2)
d = tibble::tibble(
  source = factor(rep(c("s1", "s2","s3"), each = 6L)),
  y_row = factor(rep(c("d1", "d2","d3",
                       "d1", "d2","d3"), each = 3L)),
  value = c(40, 50, 60, 40, 50, 60,40, 50, 60,
            40, 50, 60, 40, 50, 60,40, 50, 60),
)

ggplot(d) +
  aes(x=value, y=y_row) +
  geom_point(aes(group=y_row,col=source)) +
  geom_line(aes(col=source,linetype=source))+
  geom_line(aes(col=source,linetype=source),orientation = "y",
            position= position_dodge(width=0.4) )

^{Created on 2025-10-31 with reprex v2.1.1}

[DanChaltiel]

@smouksassi thank you very much, it was almost the solution!
I still had to construct a group that holds both source and y_row.
I edited the original post to clarify the issue.
Here is the final plot:

library(tidyverse)
d = tibble::tibble(
  y_row = rep(c("d1", "d2"), each = 9L),
  source = rep(rep(c("s1", "s2", "s3"), 2), rep(4:2, 2)),
  value = c(1, 2, 3, 4, 2, 3, 4, 3, 4, 2, 3, 4, 5, 3, 4, 5, 4, 5),
)
pd = position_dodge(width=0.4, orientation="y")
d %>% 
  mutate(group2 = paste(source, y_row)) %>% 
  ggplot() +
  aes(x=value, y=y_row, color=source) +
  geom_point(aes(group=source), position=pd) +
  geom_line(aes(group=group2), orientation = "y", position=pd)

^{Created on 2025-11-01 with reprex v2.1.1}

@ggplot2_team: Feel free to close this issue, but I think this would be a nice example for geom_line()

[teunbrand]

Thank you Samer for finding a solution to the issue.

Feel free to close this issue

I'll take you up on this 😁