s-news
[Top] [All Lists]

Re: random samples

To: "Roberts, J. Kyle" <jkrobert@bcm.tmc.edu>, Richard Pugh <rpugh@mango-solutions.com>, s-news@lists.biostat.wustl.edu
Subject: Re: random samples
From: "Richard M. Heiberger" <rmh@temple.edu>
Date: Sat, 22 Apr 2006 17:15:49 -0400
Think matrix!

This version is the same as I sent yesterday plus I constructed
result3 to give two columns, the first is the vector of 900 numbers
you requested and the second is the group labels

## construct data
result <- matrix(rnorm(30*30, mean=50,
                       sd=seq(1, 10, length=30)),
                 30, 30, byrow=TRUE)

print(result, digits=1)  ## verify standard deviations
apply(result, 2, var)^.5

## sort each of the 30 groups
result2 <- apply(result, 2, sort)

apply(result2, 2, var)^.5  ## verify standard deviations
print(result2, digits=1)

## rearrange into a single vector with a column of group labels
result3 <- cbind(y=as.vector(result2), group=as.vector(col(result2)))

result3





As an aside, you can start a sequence with a zero-length vector.
x <- numeric(0)
x
x <- append(x, 1:5)
x
x <- append(x, 1:5)
x

<Prev in Thread] Current Thread [Next in Thread>