R help - May 2010 - Find the three best values in every row

Hello,

   i have a dataframe with the GDP for different Country (in the columns) and
   Years (in the rows).

   Now i want for every year the best three values, if possible with name of
   the countries (columnnames).

   For the best it's no problem but for the other two values.

   Thanks,

   Alfred

Hmm...

 set.seed(17*11)
 d<-data.frame(africa=sample(50, 10), 
 		europe= sample(50, 10),
 		n.america= sample(50, 10),
 		s.america= sample(50, 10),
 		antarctica= sample((1:50)/20, 10)
 		)

#Get three top from each row
t(apply(d,1,function(x,n) x[which(rank(x)> length(x)-n)], n=3))

#Get them with names
get.n<- function(x, names, n)	{
	rmin<- 
	rv<-x[w<-which(rank(x)> length(x)-n)]
	names(rv)<-names[w]
	return(as.data.frame(rv))
}

apply(d,1,get.n, n=3, names=names(d))


>>> Alfred Schulze <Alfred-Schulze at web.de> 05/07/10 2:43 PM
>>>
   Hello,

   i have a dataframe with the GDP for different Country (in the
columns) and
   Years (in the rows).

   Now i want for every year the best three values, if possible with
name of
   the countries (columnnames).

   For the best it's no problem but for the other two values.

   Thanks,

   Alfred
______________________________________________
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.


*******************************************************************
This email and any attachments are confidential. Any use...{{dropped:8}}

Hello Alfred,

I found the solution from S. Ellison
(https://stat.ethz.ch/pipermail/r-help/2010-May/238158.html) really 
inspiring.
Here I am using tail and the library 'plyr':

  set.seed(17*11)
  d<-data.frame(africa=sample(50, 10),
  		europe= sample(50, 10),
  		n.america= sample(50, 10),
  		s.america= sample(50, 10),
  		antarctica= sample((1:50)/20, 10)
  		)

# using tail
t(apply(d, 1, function(x, n) tail(sort(x), n), n=3))

lapply(split(d, rownames(d)), function(x, n) sort(x)[n:ncol(x)], n=3)

# with plyr from Hadley Wickham
library(plyr)
ldply(split(d, rownames(d)), function(x, n) sort(x)[n : ncol(x)], n=3)

HTH
Patrick

Am 07.05.2010 15:43, schrieb Alfred Schulze:
>
>     Hello,
>
>     i have a dataframe with the GDP for different Country (in the columns)
and
>     Years (in the rows).
>
>     Now i want for every year the best three values, if possible with name
of
>     the countries (columnnames).
>
>     For the best it's no problem but for the other two values.
>
>     Thanks,
>
>     Alfred
> ______________________________________________
> 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.

-----Urspr?ngliche Nachricht-----

Von: S Ellison <S.Ellison at lgc.co.uk>

Gesendet: 07.05.2010 16:44:08

An: r-help at r-project.org,Alfred-Schulze at web.de

Betreff: Re: [R] Find the three best values in every row

>Hmm...
>
> set.seed(17*11)
> d<-data.frame(africa=sample(50, 10), 
> 		europe= sample(50, 10),
> 		n.america= sample(50, 10),
> 		s.america= sample(50, 10),
> 		antarctica= sample((1:50)/20, 10)
> 		)
>
>#Get three top from each row
>t(apply(d,1,function(x,n) x[which(rank(x)> length(x)-n)], n=3))
>
>#Get them with names
>get.n<- function(x, names, n)	{
>	rmin<- 
>	rv<-x[w<-which(rank(x)> length(x)-n)]
>	names(rv)<-names[w]
>	return(as.data.frame(rv))
>}
>
>apply(d,1,get.n, n=3, names=names(d))
>
>
>
>>>> Alfred Schulze [ 05/07/10 2:43 PM >>>
>
>   Hello,
>
>   i have a dataframe with the GDP for different Country (in the
>columns) and
>   Years (in the rows).
>
>   Now i want for every year the best three values, if possible with
>name of
>   the countries (columnnames).
>
>   For the best it's no problem but for the other two values.
>
>   Thanks,
>
>   Alfred
>______________________________________________
>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.
>
>
>*******************************************************************
>This email and any attachments are confidential. Any use, copying or
>disclosure other than by the intended recipient is unauthorised. If 
>you have received this message in error, please notify the sender 
>immediately via +44(0)20 8943 7000 or notify postmaster at lgc.co.uk 
>and delete this message and any copies from your computer and network. 
>LGC Limited. Registered in England 2991879. 
>Registered office: Queens Road, Teddington, Middlesex, TW11 0LY, UKtt



Thanks, that's what I need.

But I have some NA's in the data and 

t(apply(data.frame,1,function(x,n) x[which(rank(x)> length(x)-n)], n=3,
na.rm=TRUE))

isn't working. 


And I have a further question. How I get the value also sorted descending?

Thanks,

Alfred
___________________________________________________________
GRATIS: Movie-Flat mit ?ber 300 Top-Videos. F?r WEB.DE Nutzer
dauerhaft kostenlos! Jetzt freischalten unter http://movieflat.web.de