Order x-axis categorical variable using another categorical variable-CodePudding

I have a dataframe like follows:

Class       ID   Stage   Abundance          Substrat
OTUA      rep1  A-X1    123                  G
OTUA      rep2  A-X1    234                  PC
OTUA      rep3  A-X1    [numerical values]   [only two categorical values]
OTUB      rep1  A-X1
OTUA      rep1  EGG
OTUA      rep2  EGG
OTUA      rep1  EL
...       ...   ...    ...

I want to plot in geom_bar the ID of each sample by its stacked abundance for each different variable in the Class.

graph <- ggplot(data, aes(x=ID, y=Abundance, fill=Class))  
  facet_grid(~substrat, scales="free_x")  
  geom_bar(aes(color=Class, fill=Class), stat="identity", position="stack")

I get something like this:

this works well, but I want to order the x axis data (ID) by the stage variable, with the specific order : "A-X0", "EGG", "EL", "LL","PP","P","A-X1", but I do not want to have the stage as label, I want to still have the ID as labels.

I have tried:

graph$stage <- as.character(graph$stage)
graph$stage <- factor(x = graph$stage, levels = c("A-X0", "EGG", "EL", "LL","PP","P","A-X1"))

and:

test <- ggplot(data_ps.sam.env.rel, aes(x=reorder(ID, stage), y=Abundance, fill=Class))  
  facet_grid(~substrat, scales="free_x")  
  geom_bar(aes(color=Class, fill=Class), stat="identity", position="stack")

but i still get the samples in a random order, instead I expect that all the ID with "J8" would be together, all those with "J21" together, etc

CodePudding user response：

It could have been greate to have an example of your data using dput. What do you think about adding a new variable using paste0(d$stage, d$ID) using this new variable in ggplot, and then removing the d$stage in the label using regex?

CodePudding user response：

I resolved this issue if anyone is interested, after doing the graph, I specified the order of the factors:

graph$data$ID <- as.character(ps.1.sample_class$data$ID)
graph$data$ID <- factor(x = graph$data$ID, 
                        levels = unique(graph$data$ID[order(as.character(graph$data$stage))]))