set x-axis display range - r

Question

The example data:
nltt <- structure(list(time = c(0, 1.02579504946471, 1.66430039972234,
1.67435173974229, 1.82009140034063, 1.95574135900067, 2.06963147976273,
2.64869249896209, 3.10438864450297, 0, 0.56927280073675, 1.94864234867127,
3.40490224199942, 0, 0.444318793403606, 1.34697160089298, 5.86547288923207,
0, 1.10824151737219, 1.77801220982399, 1.82246583876729, 2.18034182214015,
2.33051663760657, 3.01615794541527, 0, 0.101501512884473, 0.98261402255534,
1.04492141817475, 1.16554239437694, 1.25441082400256, 1.25777371029976,
1.62464049949719, 1.87253384965378, 1.91118229908154, 1.94105777022533,
2.17755553127212, 2.37899716574036, 2.85003451051712, 3.16711386665322
), num = c(2, 3, 4, 5, 4, 5, 6, 5, 6, 2, 3, 4, 5, 2, 3, 2, 3,
2, 3, 4, 5, 6, 7, 6, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 10, 11,
12, 13, 14), rep = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L,
2L, 2L, 2L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L,
5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L)), row.names = c(NA,
-39L), class = "data.frame")
The example code:
ggplot2::ggplot(nltt, aes(time, num, group = as.factor(rep), color = as.factor(rep))) +
ggplot2::geom_line() + ggplot2::coord_cartesian(xlim = c(0, 3)) +
ggplot2::theme(legend.position = "none") + ggplot2::xlab("age")
The example plot:
I would like each line in the plot to stop precisely at x = 3, but adding coord_cartesian(xlim = c(0, 3)) does not achieve my goal because the lines continues in the right padding area. How can I limit the lines in the range of [0, 3] without truncating my raw data?

Answer 1

Up front: two answers, the first removes the margins and requires no change to the data; the second preserves the margins, which requires one to modify the data in-place.
Remove the margin(s)
The default behavior is to expand the margins a little. While there is the ggplot2::expansion to control the multiplicative and additive components of the expansion, it can only be used in scale_continuous which, as you know, will result in loss (NA) of out-of-bound points.
If you can accept losing the boundary on the left as well, though, you can add expand=FALSE to your coord_cartesian and get your desired results:
ggplot2::ggplot(nltt, aes(time, num, group = as.factor(rep), color = as.factor(rep))) +
ggplot2::geom_line() +
ggplot2::coord_cartesian(xlim = c(0, 3), expand = FALSE) +
ggplot2::theme(legend.position = "none") +
ggplot2::xlab("age")
If you want to retain the left margin, though, you can force it by adjusting the xlim=, realizing that the default is around expansion(mult=0.05, add=0):
ggplot2::ggplot(nltt, aes(time, num, group = as.factor(rep), color = as.factor(rep))) +
ggplot2::geom_line() +
ggplot2::coord_cartesian(xlim = c(-0.15, 3), expand = FALSE) +
ggplot2::theme(legend.position = "none") +
ggplot2::xlab("age")
Interpolate and truncate, external to ggplot
ggplot2::scale_x_continuous(..., oob=) supports several mechanisms for dealing with out-of-bounds data, including:
the default censor (replaces with NA), which doesn't work since we don't have data at time=3
scales::squish that will take (for example) x=4 and squish it back to x=3; the unfortunate side-effect of this is that it is univariate (it does not attempt to change the corresponding y= value), so the slopes of the squished line segments will be steeper, and (in my mind at least) this corrupts the data and vis;
a user-defined function that is passed the values and the associated limits; unfortunately, it is also univariate, so we're stuck with the same data/vis slope-corruption as the previous bullet.
This brings me to the suggestion to interpolate the data yourself before passing to ggplot. I'll demo with dplyr but it can be done easily with base R or other dialects as well.
library(dplyr)
group_by(nltt, rep) %>%
## step 1: interpolate, returns *just* time=3 data, nothing more
summarize(as.data.frame(setNames(approx(time, num, xout = 3), c("time", "num")))) %>%
## step 2: combine with the original data
bind_rows(nltt) %>%
## step 3: remove data over 3
dplyr::filter(time <= 3) %>%
ggplot(aes(time, num, group = as.factor(rep), color = as.factor(rep))) +
ggplot2::geom_line() + ggplot2::coord_cartesian(xlim = c(0, 3)) +
ggplot2::theme(legend.position = "none") +
ggplot2::xlab("age")