[R] Squaring one column of data

[baptiste auguie]

Kenny Larsen wrote:
> Hi,
>
> A fairly basic problem I think, here although searching the inetrnet doesn't
> seem to reveal a solution. I have a dataset with two columns of real
> numbers. It is read in via read.table. I simply need to square the data in
> the second column and then plot it. I have tried extracting the second
> column (b <- z[2]) and then squaring it and merging it to a variable which
> is the first column extracted. Firstly thsi seems very convoluted, secondly
> it gives a seg fault (I'm guessing because memory hasn't been allocated for
> the second column on the creation of teh first variable).
>
> So any ideas about the easy way of doing the matsh so I can then just use
> plot(z, .....) as before?
>
> Cheers in advance,
>
> Kenny
>   

Hi,

A few alternatives,

d <- data.frame(x = seq(1, 10), y=seq(1, 10))
d2 <- d # copy

# variants that make a new column
d$z <- d$y^2
d <- within(d, {z1 <- y^2})

# destructive transformation
d3 <- transform(d2, y=y^2)

ize  <-  # this may be useful if you want to apply a function to several 
columns (orig. idea from vQ)
function(d, columns=names(d), izer=`^`, ...)
{
   d[columns] = sapply(d[columns], izer, ...)
   d
}

d4 <- ize(d2, "y",`^`, 2)


HTH,

baptiste