[R] Sampling with conditions

[Dennis Murphy]

In addition to Dan's quite valid concern,  the final sample is not
truly 'random' - the first k - 1 elements are randomly chosen, but the
last is determined so that the constraint is met.

Dennis

On Tue, Nov 8, 2011 at 9:59 AM, Nordlund, Dan (DSHS/RDA)
<NordlDJ at dshs.wa.gov> wrote:
>> -----Original Message-----
>> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
>> project.org] On Behalf Of SarahJoyes
>> Sent: Tuesday, November 08, 2011 5:57 AM
>> To: r-help at r-project.org
>> Subject: Re: [R] Sampling with conditions
>>
>> That is exactly what I want, and it's so simple!
>> Thanks so much!
>>
>
> Sarah,
>
> I want to point out that my post was qualified by "something like".  I am not sure it is exactly what you want.  Since you didn't quote my post, let me show my suggestion and then express my concern.
>
> n <- matrix(0,nrow=5, ncol=10)
> repeat{
>  c1 <- sample(0:10, 4, replace=TRUE)
>  if(sum(c1) <= 10) break
> }
> n[,1] <- c(c1,10-sum(c1))
> n
>
> This nominally meets your criteria, but it will tend to result in larger digits being under-represented.  For example, you unlikely to get a result like c(0,8,0,0,2) or (9,0,0,1,0).
>
> That may be OK for your purposes, but I wanted to point it out.
>
> You could use "something like"
>
> n <- matrix(0,nrow=5, ncol=10)
> c1 <- rep(0,4)
> for(i in 1:4){
>  upper <- 10-sum(c1)
>  c1[i] <- sample(0:upper, 1, replace=TRUE)
>  if(sum(c1) == 10) break
> }
> n[,1] <- c(c1,10-sum(c1))
> n
>
> if that would suit your purposes better.
>
>
> Good luck,
>
> Dan
>
> Daniel J. Nordlund
> Washington State Department of Social and Health Services
> Planning, Performance, and Accountability
> Research and Data Analysis Division
> Olympia, WA 98504-5204
>
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>