ggplot2 + loop results in "Error: Discrete value supplied to continuous scale" - loops

Question

I've been fumbling with this for a while now but let me start with the beginning.
This is how the first few lines of my data look like:
Year Action_mean_global Adventure_mean_global Fighting_mean_global Misc_mean_global Platform_mean_global
1 1980 0.3400000 NaN 0.77 0.6775 NaN
2 1981 0.5936000 NaN NaN NaN 2.3100
3 1982 0.3622222 NaN NaN 0.8700 1.0060
4 1983 0.4085714 0.4 NaN 2.1400 1.3860
5 1984 1.8500000 NaN NaN 1.4500 0.6900
6 1985 1.7600000 NaN 1.05 NaN 10.7925
Puzzle_mean_global Racing_mean_global Roleplaying_mean_global Shooter_mean_global Simulation_mean_global
1 NaN NaN NaN 3.53500 NaN
2 1.120000 0.480000 NaN 1.00400 0.45
3 3.343333 0.785000 NaN 0.75800 NaN
4 0.780000 NaN NaN 0.48000 NaN
5 1.046667 1.983333 NaN 10.36667 NaN
6 0.802500 NaN NaN 1.00000 0.03
Sports_mean_global Strategy_mean_global Total_mean_global
1 0.4900 NaN 1.2644444
2 0.1975 NaN 0.7776087
3 0.5250 NaN 0.8016667
4 3.2000 NaN 0.9876471
5 3.0900 NaN 3.5971429
6 1.9600 NaN 3.8528571
They are all numeric.
Now, I simply wanted to do a plot with ggplot() + geom_line() to visualize change over year per genre. It works when doing it step by step, i.e.:
ggplot(df)+
geom_line(aes_string(x = 'Year', y = plot_vector[1]))
geom_line(aes_string(x = 'Year', y = plot_vector[2]))+
geom_line(aes_string(x = 'Year', y = plot_vector[3]))+
geom_line(aes_string(x = 'Year', y = plot_vector[4]))+
geom_line(aes_string(x = 'Year', y = plot_vector[5]))+
geom_line(aes_string(x = 'Year', y = plot_vector[6]))+
geom_line(aes_string(x = 'Year', y = plot_vector[7]))+
geom_line(aes_string(x = 'Year', y = plot_vector[8]))+
geom_line(aes_string(x = 'Year', y = plot_vector[9]))+
geom_line(aes_string(x = 'Year', y = plot_vector[10]))+
geom_line(aes_string(x = 'Year', y = plot_vector[11]))+
geom_line(aes_string(x = 'Year', y = plot_vector[12]))
(plot_vector simply contains all column-names except for Year)
However, doing it in a for-loop:
p1 <- ggplot(df)+
geom_line(aes_string(x = 'Year', y = plot_vector[1]))
for (plotnumber in 2:length(plot_vector))
{
p1 <- p1 +
geom_line(aes_string(x = 'Year', y = plot_vector[plotnumber]))
}
I get the error message. Anyone can muster an idea?

Answer 1

Adding lines with a for loop to a ggplot object may have caused the reported error message but has a general problem caused by lazy evaluation. This has been asked frequently on SO, see, e.g., “for” loop only adds the final ggplot layer, or ggplot loop adding curves fails, but works one at a time.
However, ggplot2 works best when data are supplied in long format. Here, melt()from the data.table package is used to reshape df:
library(data.table)
molten <- melt(setDT(df), id.vars = c("Year"))
library(ggplot2)
ggplot(molten, aes(x = Year, y = value, group = variable, colour = variable)) +
geom_line()
This creates the following chart:
Data
df <- structure(list(Year = 1980:1985, Action_mean_global = c(0.34,
0.5936, 0.3622222, 0.4085714, 1.85, 1.76), Adventure_mean_global = c(NaN,
NaN, NaN, 0.4, NaN, NaN), Fighting_mean_global = c(0.77, NaN,
NaN, NaN, NaN, 1.05), Misc_mean_global = c(0.6775, NaN, 0.87,
2.14, 1.45, NaN), Platform_mean_global = c(NaN, 2.31, 1.006,
1.386, 0.69, 10.7925), Puzzle_mean_global = c(NaN, 1.12, 3.343333,
0.78, 1.046667, 0.8025), Racing_mean_global = c(NaN, 0.48, 0.785,
NaN, 1.983333, NaN), Roleplaying_mean_global = c(NaN, NaN, NaN,
NaN, NaN, NaN), Shooter_mean_global = c(3.535, 1.004, 0.758,
0.48, 10.36667, 1), Simulation_mean_global = c(NaN, 0.45, NaN,
NaN, NaN, 0.03), Sports_mean_global = c(0.49, 0.1975, 0.525,
3.2, 3.09, 1.96), Strategy_mean_global = c(NaN, NaN, NaN, NaN,
NaN, NaN), Total_mean_global = c(1.2644444, 0.7776087, 0.8016667,
0.9876471, 3.5971429, 3.8528571)), .Names = c("Year", "Action_mean_global",
"Adventure_mean_global", "Fighting_mean_global", "Misc_mean_global",
"Platform_mean_global", "Puzzle_mean_global", "Racing_mean_global",
"Roleplaying_mean_global", "Shooter_mean_global", "Simulation_mean_global",
"Sports_mean_global", "Strategy_mean_global", "Total_mean_global"
), class = "data.frame", row.names = c(NA, -6L))