Sort a data frame when column name is assigned in a variable

Question

I am new to R and hence my question is likely to be basic. I tried researching the answer before posting here however didn't get the answer I was looking for.

I am trying to order a data frame on the basis of certain columns. The column is dynamic and depends on a certain "outcome" which would be from a vector of "validoutcomes".

validoutcomes <- c("A","B","C")

if outcome is "A" then I am supposed to read data from column 11. For "B" its columns 17 and for "C" it is 23.

I renamed the columns so that it is easier for my readability and alignment with validoutcomes. "input" is my source data from the CSV.

colnames(input)[11] <- validoutcomes[1] colnames(input)[17] <- validoutcomes[2] colnames(input)[23] <- validoutcomes[3]

I then tried to sort using

statelist <- statelist[order(statelist$outcome)]

Any ideas?

Answer 1

Your first attempt was more accurate. Since I do not have your data I will just use the iris data set,

 iris_ordered_on_Sepal_Length <- iris[order(iris$Sepal.Length),]

or if you use data.table,

iris_dt <- data.table(iris)
iris_dt[order(Sepal.Length)]

data.table will make renaming easier too,

setnames(input, c("colname_11, "colname_17", "colname_23"), validoutcomes[1 : 3])

Answer 2

Ok, I have found the likely answer. I am not sure if this is the best answer but it worked for me.

Since $ is used to call the colnames, R was looking for a column name as outcome. However all it needed was square brackets[]. I replaced it and so far seems to be working for me.

statelist <- statelist[order(statelist[outcome])]

I am still very interested in knowing if there is a better solution or if someone sees issue with my observation.