Suppress separator in paste when values are missing

Question

This is my data structure that I read into a dataframe.

treatment  egf       mean      se
10 uM PP2            -697.25   14124.349
10 uM PP2  1 nM EGF  14715.50  8862.012
DMSO                 58589.25  7204.824
DMSO       1 nM EGF  87852.00  12149.159

The combination of the treatment and egf columns represent the unique id for each column. I would like to create a column that combines these so that I can have one column that uniquely represents each row. However, because of the missing values in the EGF column, when I use paste, it does this annoying thing:

>paste(rawp$treatment, rawp$egf, sep=" + ")
[1] "10 uM PP2 + "         "10 uM PP2 + 1 nM EGF" "DMSO + "             
[4] "DMSO + 1 nM EGF"

where it will still place the separator there when the value is missing. I would like for it to read:

[1] "10 uM PP2"         "10 uM PP2 + 1 nM EGF" "DMSO"             
[4] "DMSO + 1 nM EGF"

How can I do this?

The whole reason I want to do this is because I want to plot the data with ggplot and it seems it requires only 1 unique column when specifying the x axis.

ggplot(data=rawp, aes(x=treatment, y=mean)) + geom_bar(stat="identity")

So if you also know of an alternative way to use combined columns to specify the category on the x axis, that would be helpful.

Answer 1

It would have been nice to use sep = ifelse(egf == "", "", " + ") but the sep argument to paste cannot be a vector. The solution: pass it as a regular argument!

with(rawp, paste0(treatment, ifelse(egf == "", "", " + "), egf))