R help - Jul 2011 - how to replace values in x by means in subgroups created in ... (not loops)

# Dear all,
# how to replace values in x by means in subgroups created in ...
# replace only these values where y=0 in continous sequence
# replace by mean calculated locally for each subgroup created by 
# continous sequence of 0,0,0 in parallel y vector, i.e. 
# where there is continous sequence of 0 in data frame vector y 
#    but we do not replace values in x[i], if y[i]!=0 
# we do not want use loops we do not use apply (not very fast) 


x      <-c(0 ,1,2,3,4,5,6,7,8,NA,NA,1  ,1 ,NA,2  ,2)
y      <-c(0 ,0,0,1,0,0,0,1,1,1 ,NA,0  ,0 ,0 ,0  ,1)
Must_be<-c(1 ,1,1,3,5,5,5,7,8,NA,NA,1.5,1 ,NA,1.5,2)

(df<-as.data.frame(cbind(x,y))  )

# I have traied many bad colusions based on cumsum, pmin, pmax, ... 
(mean_dfx_if_yIs0<-y*cumsum(x*y)/(cumsum(y)*y) )

# how to do this?
# thans for any advice 
# E

	[[alternative HTML version deleted]]

Something like this?

dataset <- data.frame(x = x, y = y)
dataset$Group <- cumsum(c(0, diff(!is.na(dataset$y) & dataset$y == 0)) ==
1)
library(plyr)
tmp <- ddply(subset(dataset, y == 0), .(Group), function(z){c(Mean =
mean(z$x, na.rm = TRUE))})
result <- merge(dataset, tmp)
result$Mean[is.na(result$y) | result$y != 0] <- result$x[is.na(result$y) |
result$y != 0]

Best regards,

Thierry
> -----Oorspronkelijk bericht-----
> Van: r-help-bounces at r-project.org [mailto:r-help-bounces at
r-project.org]
> Namens Eugeniusz Kaluza
> Verzonden: woensdag 27 juli 2011 11:21
> Aan: r-help at r-project.org
> Onderwerp: [R] how to replace values in x by means in subgroups created in
...
> (not loops)
> 
> 
> # Dear all,
> # how to replace values in x by means in subgroups created in ...
> # replace only these values where y=0 in continous sequence # replace by
mean
> calculated locally for each subgroup created by # continous sequence of
0,0,0 in
> parallel y vector, i.e.
> # where there is continous sequence of 0 in data frame vector y
> #    but we do not replace values in x[i], if y[i]!=0
> # we do not want use loops we do not use apply (not very fast)
> 
> 
> x      <-c(0 ,1,2,3,4,5,6,7,8,NA,NA,1  ,1 ,NA,2  ,2)
> y      <-c(0 ,0,0,1,0,0,0,1,1,1 ,NA,0  ,0 ,0 ,0  ,1)
> Must_be<-c(1 ,1,1,3,5,5,5,7,8,NA,NA,1.5,1 ,NA,1.5,2)
> 
> (df<-as.data.frame(cbind(x,y))  )
> 
> # I have traied many bad colusions based on cumsum, pmin, pmax, ...
> (mean_dfx_if_yIs0<-y*cumsum(x*y)/(cumsum(y)*y) )
> 
> # how to do this?
> # thans for any advice
> # E
> 
> 	[[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.