0
votes
df <- data.frame(values = c(2.5,12,4.8,56,78),samples = c('45fe.K2','59ji.K2','59rc.K1','45hi.K1','96hu.K1'),group = c('K2','K2','K1','K1','K1'))

 df
  values samples group
1    2.5 45fe.K2    K2
2   12.0 59ji.K2    K2
3    4.8 59rc.K1    K1
4   56.0 45hi.K1    K1
5   78.0 96hu.K1    K1

I want to generate a groupgrouped boxplot. So I want one plot with a K1 and K2 boxplot. I thought this https://www.r-graph-gallery.com/265-grouped-boxplot-with-ggplot2.html would do it but I can“t figure out how

p1 <- ggplot(df, aes(x=group, y=values, fill=group)) + 
    geom_boxplot() +
    facet_wrap(~group)

What can I do about that? I tried also x=samplesbut that is wrong.

EDIT: Maybe that is another question. But when I add the group column with the following code the great answer by @rodolfoksveiga results in an error

df <- data.frame(values = c(2.5, 12, 4.8, 56, 78),
samples = c('45fe.K2', '59ji.K2', '59rc.K1', '45hi.K1', '96hu.K1'))

df$group <- NA
df$group <- apply(df,1,function(x)
{ifelse(grepl('K2',df$samples) == TRUE,paste('K2'),paste('K1'))})

Error in `$<-.data.frame`(`*tmp*`, "PANEL", value = c(1L, 2L, 2L, 2L,  : 
 replacement has ... rows, data has ....rows
1
Maybe is this what you are looking for: ggplot(df, aes(x=group, y=values, fill=samples)) + geom_boxplot() - Duck
@Duck thanks, but I am getting the same Error as with my code above <error/rlang_error> Aesthetics must be either length 1 or the same as the data (5): x - JORIS
@JORIS the code that Duck suggested is working with the data that you provided. - Werner

1 Answers

0
votes

Welcome to Stack Overflow Joris.

Maybe this is what you want:

library(ggplot2)
df <- data.frame(values = c(2.5, 12, 4.8, 56, 78),
                 samples = c('45fe.K2', '59ji.K2', '59rc.K1', '45hi.K1', '96hu.K1'),
                 group = c('K2', 'K2', 'K1', 'K1', 'K1'))
ggplot(df, aes(x = group, y = values, fill = group)) + 
  geom_boxplot() +
  facet_wrap(. ~ group, scales = 'free_x')

Here is the output:

enter image description here

Let us know if this is what you're looking for.