R help - Oct 2010 - vectorizing: selecting one record per group

Hi,

 

I want to select a subsample from my data, choosing one record from each
group. I know how to do this with a for.

 

For example: lets say I have the data:

A=cbind(rnorm(100),runif(100),(rep(c(1,2,3,4,5),20)))

Where the third column is the group variable. Then what I want is to select
5 observations. Each one taken randomly from each group.

 

 

INDEX =NULL

i=1

for(index_g in  unique(A[,3])){

INDEX [i]=sample(which(A[,3]==index_g),1)

i=i+1

}

SEL=A[INDEX,]

 

 

Is there a way to do this without a “for”? in other words is there a way to
“vectorize” this?

 

Thank you,

 

 

Mauricio Romero 

Quantil S.A.S.

Bogotá,Colombia

www.quantil.com.co

 

"It is from the earth that we must find our substance; it is on the earth
that we must find solutions to the problems that promise to destroy all life
here"

 


	[[alternative HTML version deleted]]

Hello,

There are probably many ways to do this, but I think
it's easier if you use a data.frame as your object.

The easy solution for the matrix you provide is escaping
me at the moment.

One solution, using the plyr package:


library(plyr)
A <- data.frame(a = rnorm(100),b = runif(100), c = rep(c(1,2,3,4,5),20))
ddply(A, .(c), function(x) x[sample(1:nrow(x), 1), ])

             a         b c
1  0.02995847 0.4763819 1
2  0.72035194 0.2948611 2
3  1.34963917 0.2057488 3
4 -1.99427160 0.1147923 4
5 -0.73612703 0.5889539 5


Mauricio Romero wrote:
> Hi,
> 
>  
> 
> I want to select a subsample from my data, choosing one record from each
> group. I know how to do this with a for.
> 
>  
> 
> For example: lets say I have the data:
> 
> A=cbind(rnorm(100),runif(100),(rep(c(1,2,3,4,5),20)))
> 
> Where the third column is the group variable. Then what I want is to select
> 5 observations. Each one taken randomly from each group.
> 
>  
> 
>  
> 
> INDEX =NULL
> 
> i=1
> 
> for(index_g in  unique(A[,3])){
> 
> INDEX [i]=sample(which(A[,3]==index_g),1)
> 
> i=i+1
> 
> }
> 
> SEL=A[INDEX,]
> 
>  
> 
>  
> 
> Is there a way to do this without a ?for?? in other words is there a way to
> ?vectorize? this?
> 
>  
> 
> Thank you,
> 
>  
> 
>  
> 
> Mauricio Romero 
> 
> Quantil S.A.S.
> 
> Bogot?,Colombia
> 
> www.quantil.com.co
> 
>  
> 
> "It is from the earth that we must find our substance; it is on the
earth
> that we must find solutions to the problems that promise to destroy all
life
> here"
> 
>  
> 
> 
> 	[[alternative HTML version deleted]]
> 
> 
> 
> ------------------------------------------------------------------------
> 
> ______________________________________________
> 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.

Try this:

A <- data.frame(V1 = rnorm(100), V2 = runif(100), V3 = rep(c(1,2,3,4,5),20))
do.call(cbind, lapply(aggregate(. ~ V3, A, FUN = sample, size = 5), c))

On Wed, Oct 13, 2010 at 5:01 PM, Mauricio Romero <
mauricio.romero@quantil.com.co> wrote:
> Hi,
>
>
>
> I want to select a subsample from my data, choosing one record from each
> group. I know how to do this with a for.
>
>
>
> For example: lets say I have the data:
>
> A=cbind(rnorm(100),runif(100),(rep(c(1,2,3,4,5),20)))
>
> Where the third column is the group variable. Then what I want is to select
> 5 observations. Each one taken randomly from each group.
>
>
>
>
>
> INDEX =NULL
>
> i=1
>
> for(index_g in  unique(A[,3])){
>
> INDEX [i]=sample(which(A[,3]==index_g),1)
>
> i=i+1
>
> }
>
> SEL=A[INDEX,]
>
>
>
>
>
> Is there a way to do this without a “for”? in other words is there a way to
> “vectorize” this?
>
>
>
> Thank you,
>
>
>
>
>
> Mauricio Romero
>
> Quantil S.A.S.
>
> Bogotá,Colombia
>
> www.quantil.com.co
>
>
>
> "It is from the earth that we must find our substance; it is on the
earth
> that we must find solutions to the problems that promise to destroy all
> life
> here"
>
>
>
>
>        [[alternative HTML version deleted]]
>
>
> ______________________________________________
> R-help@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.
>
>

-- 
Henrique Dallazuanna
Curitiba-Paraná-Brasil
25° 25' 40" S 49° 16' 22" O

	[[alternative HTML version deleted]]