You can also do this:
set.seed(28)
t1<-data.frame(id=rep(1:3,rep(3,3)),dt=rep(1:3,rep(9,3)),var=c('num1','num2','norm'),value=rnorm(27))
t2<-t1[t1$var=="norm",]
t3<-t1[t1$var!="norm",]
df1<-merge(t2,t3,by=c("id","dt"))
df1$Norm<-df1$value.y/df1$value.x
df2<-df1[,c(1:2,5,7)]
colnames(df2)[3]<- "var1"
head(df2)
# id dt var1 Norm
#1 1 1 num1 1.42893951
#2 1 1 num2 0.04829956
#3 1 2 num1 1.21120856
#4 1 2 num2 -0.43157114
#5 1 3 num1 -0.74132293
#6 1 3 num2 -1.99100489
#or to preserve the order
library(plyr)
dfNew<- join(t2,t3,by=c("id","dt"))
dfNew$Norm<- dfNew[,6]/dfNew[,4]
dfNew<- dfNew[,c(1,2,5,7)]
head(dfNew)
# id dt var Norm
#1 1 1 num1 1.42893951
#2 1 1 num2 0.04829956
#3 2 1 num1 -3.42492323
#4 2 1 num2 0.30611744
#5 3 1 num1 -0.94940980
#6 3 1 num2 -0.06622304
identical(res,dfNew)
#[1] TRUE
A.K.
----- Original Message -----
From: arun <smartpink...@yahoo.com>
To: neal subscribe <nealsubscr...@gmail.com>
Cc: R help <r-help@r-project.org>
Sent: Tuesday, April 9, 2013 12:59 AM
Subject: Re: [R] (no subject)
Hi,
Try this:
set.seed(28)
t1<-data.frame(id=rep(1:3,rep(3,3)),dt=rep(1:3,rep(9,3)),var=c('num1','num2','norm'),value=rnorm(27))
head(t1)
# id dt var value
#1 1 1 num1 -1.90215722
#2 1 1 num2 -0.06429479
#3 1 1 norm -1.33116707
#4 2 1 num1 -1.81999167
#5 2 1 num2 0.16266969
#6 2 1 norm 0.53139634
res<-do.call(rbind,lapply(split(t1,list(t1$id,t1$dt)),function(x)
{x$Norm<-x$value/tail(x$value,1);head(x[,-4],-1)}))
row.names(res)<-1:nrow(res)
head(res)
# id dt var Norm
#1 1 1 num1 1.42893951
#2 1 1 num2 0.04829956
#3 2 1 num1 -3.42492323
#4 2 1 num2 0.30611744
#5 3 1 num1 -0.94940980
#6 3 1 num2 -0.06622304
A.K.
----- Original Message -----
From: neal subscribe <nealsubscr...@gmail.com>
To: r-help@r-project.org
Cc:
Sent: Monday, April 8, 2013 10:49 PM
Subject: [R] (no subject)
Hi
I would like to normalize my data by one of the variables in long format.
My data is like this:
>
t1<-data.frame(id=rep(1:3,rep(3,3)),dt=rep(1:3,rep(9,3)),var=c('num1','num2','norm'),value=rnorm(27))
> t1
id dt var value
1 1 1 num1 -1.83276256
2 1 1 num2 1.57034303
3 1 1 norm 0.60008563
4 2 1 num1 -0.96893477
5 2 1 num2 0.30423346
6 2 1 norm -0.07044640
7 3 1 num1 -1.30558219
...
in this case there are 3 ids and 3 dates (dt). on each date, each id has a
normalization factor given by the variable norm. i would like to divide all
the variables for each id by the normalization on that date. (i don't care
if the normalization is divided by itself since it should always be end up
1, might be a useful check.) so i would like to end up with a data frame
like
1 1 num1 -1.83/.6
1 1 num2 -1.57/.6
2 1 num1 -.97/(-.07)
2 1 num2 .3/(-.07)
etc. The actual data frame is quite long with about many id's and dates.
Thanks!
Neal
[[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.
______________________________________________
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.
