[R] variable substitution

[Erik Iverson]

Hello,

Jon Erik Ween wrote:
> Hi
> 
> I would like to write a script that reads a list of variable names. These variable names are some of the column headers in a data.frame. Then I want do a for-loop to execute various operations on the specified variables in the data.frame, but can't figure out how to do the necessary variable substitution. In bash (or C) I would use "$var", but there seems to be no equivalent in R. The data.frame has 300 columns and 2500 rows. Not all columns are continuous variables, some are factors, descriptors, etc., so I use the varlist to pick which columns I want to analyze.
> 
> #Example script
> varlist<-read.table(/path/to/varlist)
> for (i in 1:length(varlist)){
> 	res<-mean(Dataset$SOMETHINGHERE_i )
> 	write(res) somewhere
> }
> 

In your script, perhaps

res <- mean(Dataset[[varlist[i]]])

would work.

Better might be:

colMeans(Dataset[varlist])

or if your actual function is not "mean":

sapply(Dataset[varlist], function)

where "function" is whatever you want (e.g., mean, sd, ...)

and do avoid the "varlist" problem completely, say you only want to run 
it on numeric variables

sapply(Dataset[sapply(Dataset, is.numeric)], function)

None of this is tested, but the ideas should work.