[R] Equivalent to a BY command in SAS

[Rolf Turner]

On 14/04/2008, at 2:36 PM, zerfetzen wrote:

>
> Hi,
> I'm very new to R and absolutely love it.  Does anyone know how to use
> something in R that functions like a BY command in SAS?
>
> For example, let's say you have a variable x, and you want to see  
> the mean.
> Easy...
>
>> mean(x)
>
> But what if you want to see the mean of x conditional on another  
> discrete
> variable?  My best attempts so far are something like...
>
>> mean(x, y_cat=1)
>
> ...which of course doesn't work.  I have downloaded plenty of R  
> user guides
> that are very informative, but am not seeing much on detailed  
> descriptives
> or data manipulation (for my life, I can't figure out how to sort an
> attached data frame, but that's another issue).  Thanks.

 > set.seed(42)
 > x <- rnorm(100)
 > y <- sample(letters[1:3],100,TRUE)
 > by(x,y,mean)
INDICES: a
[1] 0.1089523
------------------------------------------------------------
INDICES: b
[1] -0.2253035
------------------------------------------------------------
INDICES: c
[1] 0.2997985

Or:

  > tapply(x,y,mean)
          a          b          c
  0.1089523 -0.2253035  0.2997985

		cheers,

			Rolf Turner


######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}