Question

I have a question regarding a weird result for a line plot that uses facets.
I have water data masurements for different depth (=pressures). The data comes as a table as such:

Pressure Temperature pH
0        30          8.1
1        28          8.0

I “melt” this data to yield:

Pressure variable    value
0        Temperature 30
1        Temperature 30
0        pH          8.1
1        pH          8.0

and so on. I now plot this:

ggplot(data.m.df, aes(x=value, y=Pressure)) +
  facet_grid(.~variable, scale = "free") +
  scale_y_reverse() +
  geom_line() +
  opts(axis.title.x=theme_blank())

It kinda works, except there are parts of the line plot that get filled with solid color. I have no idea why, especially because it works just fine if I exchange x for y and use “variable ~ .” as the facet_grid formula.

Answer 1

Note the difference between geom_line and geom_path applied to the same data.

library(ggplot2)

x = c(seq(1, 10, 1), seq(10, 1, -1))
y = seq(0, 19, 1)
df = data.frame(x, y)

ggplot(df, aes(x, y)) + geom_line()
ggplot(df, aes(x, y)) + geom_path() 

Note the order in the df data frame.

    x  y
1   1  0
2   2  1
3   3  2
4   4  3
5   5  4
6   6  5
7   7  6
8   8  7
9   9  8
10 10  9
11 10 10
12  9 11
13  8 12
14  7 13
15  6 14
16  5 15
17  4 16
18  3 17
19  2 18
20  1 19

geom_path plots in order of the observations.

geom_line plots in order of x values.

The effect is more marked when the x values are closer together.

x = c(seq(1, 10, .01), seq(10, 1, -.01))
y = seq(.99, 19, .01)
df = data.frame(x, y)

ggplot(df, aes(x, y)) + geom_line() 
ggplot(df, aes(x, y)) + geom_path()