[R] Converting characters to numbers in data frames

[Gabor Grothendieck]

On 8/25/05, Anon. <bob.ohara at helsinki.fi> wrote:
> I'm sure I'm missing something obvious here, but I can't find the
> solution (including in the FAQ etc.).
> 
> I have a vector of names of variables like this: NRes.x.y. where x and y
> are numbers.  I want to extract these numbers as numbers to use
> elsewhere.  I can extract the numbers as a list of characters using
> strsplit(), and convert that to a data frame, e.g.:
> 
> NAMES=c("NRes.1.2.", "NRes.1.3.", "NRes.1.4.", "NRes.1.5.", "NRes.1.6.")
> NUMBERS=strsplit(gsub("NRes.","", NAMES, perl =T), '.', fixed = TRUE)
> NUMBERS.df=t(data.frame(NUMBERS))
> 
> But I now want to convert the characters to be numeric.  Using
> as.numeric(NUMBERS.df) converts them, but to a vector.  How can I
> convert and keep as a data frame?  I could use this:
> 
> matrix(as.numeric(NUMBERS.df), ncol=dim(NUMBERS.df)[2])
> 
> but I seem to be jumping through far too many hoops: there must be an
> easier way.  An suggestions?


Try this:

   read.table(textConnection(NAMES), sep = ".")[,2:3]

You may also want to add the col.names= argument to the read.table call
if you want prettier column names.  Also I think Names and Numbers would
be sufficient to distinguish them from potential lower case counterparts.