R help - Jul 2008 - how to speed up this for loop?

Could anyone tell me a better way to achieve the output of this for loop? It
seems to run quite slow. I'm sure there must be a more consise way to sum
from FN to LN, excluding positive values, for each row.

#sum between FN and LN, excluding positive values
for(i in 1:R){
for(j in FN[i]:LN[i]){
if(Temp[i,j]<0)
sum[i] <- sum[i] + sum(Temp[i,j])}}

Cheers, 
R
-- 
View this message in context:
http://www.nabble.com/how-to-speed-up-this-for-loop--tp18579577p18579577.html
Sent from the R help mailing list archive at Nabble.com.

I'm not sure if I understand correct, but maybe something like:

apply(dat,2,function(x) sum(x[x<0]))

Please do read the posting guide and post a small, working example!!!

Bart


Rheannon wrote:
> 
> Could anyone tell me a better way to achieve the output of this for loop?
> It seems to run quite slow. I'm sure there must be a more consise way
to
> sum from FN to LN, excluding positive values, for each row.
> 
> #sum between FN and LN, excluding positive values
> for(i in 1:R){
> for(j in FN[i]:LN[i]){
> if(Temp[i,j]<0)
> sum[i] <- sum[i] + sum(Temp[i,j])}}
> 
> Cheers, 
> R
> 
-- 
View this message in context:
http://www.nabble.com/how-to-speed-up-this-for-loop--tp18579577p18584610.html
Sent from the R help mailing list archive at Nabble.com.

Hi

r-help-bounces at r-project.org napsal dne 22.07.2008 01:24:05:
> 
> Could anyone tell me a better way to achieve the output of this for 
loop? It
> seems to run quite slow. I'm sure there must be a more consise way to 
sum
> from FN to LN, excluding positive values, for each row.
It is difficult to understand what you really want as we do not have R, 
FN, LN and Temp.
> vec<-rnorm(100)
> sum(vec[vec<0])
[1] -45.72518

Gives you sum of vec values which are lower than zero. It shall be easy to 
extend it to set of vectors or matrix.
> mat<-rnorm(120)
> dim(mat)<-c(10,12)
> apply(mat, 2, function(x) sum(x[x<0]))
 [1] -0.9543727 -3.9965437 -6.7924017 -1.4949170 -3.8042263 -6.5552674 
-2.6386912 -3.3587444 -6.3740483 -3.1178593 -7.2580333 -3.2709297

There are other options but you shall follow posting guide to obtain 
better answers.

Regards
Petr

> 
> #sum between FN and LN, excluding positive values
> for(i in 1:R){
> for(j in FN[i]:LN[i]){
> if(Temp[i,j]<0)
> sum[i] <- sum[i] + sum(Temp[i,j])}}
> 
> Cheers, 
> R
> -- 
> View this message in context: 
http://www.nabble.com/how-to-speed-up-this-for-
> loop--tp18579577p18579577.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> 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.

Here is the full code:

R <- 5328
#Temp is a temperature file with daily temp. values for each day of the
year, for 5328 locations
Temp <- rnorm(100)
dim(Temp) <- c(365,5328)
# transpose temp dataframe in order to perform rollmean (running mean) on
proper data section
Temp_T <- t(Temp)
library(zoo)
Temp_T <- zoo(Temp_T)

#**** 31 DAY RUNNING MEAN *********
RM <- rollmean(Temp_T, 31)

#function that finds first negative number, if there are none a value of 365
will be output
firstneg = function(x)if(any(x<0))min(which(x<0)) else 335

#apply the function to every column in Matrix
FN <- apply(RM,2,firstneg)

#problem: column values are off by 30 (31 day running mean)
FN <- matrix(FN) 
add = function(x)x=x+30
FN <- apply(FN,2,add)

#function that finds last negative number, if there are none a value of 365
will be output
lastneg = function(x)if(any(x<0))max(which(x<0)) else 335

#apply the function to every column in Matrix
LN <- apply(RM,2,lastneg)

#problem: column values are off by 30
LN <- matrix(LN) 
LN <- apply(LN,2,add)
RMT <- t(RM)

#Create and fill an empty matrix with zeros
sum <- matrix(nrow=R,ncol=1)
for(i in 1:R){sum[i] <- 0}

#******This is the loop that I would like to speed up************

#sum values between FN and LN, excluding positive values
for(i in 1:R){
for(j in FN[i]:LN[i]){
if(Temp[i,j]<0)
sum[i] <- sum[i] + sum(Temp[i,j])}}

-- 
View this message in context:
http://www.nabble.com/how-to-speed-up-this-for-loop--tp18579577p18592932.html
Sent from the R help mailing list archive at Nabble.com.

> -----Original Message-----
> From: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] On Behalf Of Rheannon
> Sent: Tuesday, July 22, 2008 9:09 AM
> To: r-help at r-project.org
> Subject: Re: [R] how to speed up this for loop?
> 
> 
> Here is the full code:
> 
> R <- 5328
> #Temp is a temperature file with daily temp. values for each 
> day of the
> year, for 5328 locations
> Temp <- rnorm(100)
> dim(Temp) <- c(365,5328)
> # transpose temp dataframe in order to perform rollmean 
> (running mean) on
> proper data section
> Temp_T <- t(Temp)
> library(zoo)
> Temp_T <- zoo(Temp_T)
> 
> #**** 31 DAY RUNNING MEAN *********
> RM <- rollmean(Temp_T, 31)
> 
> #function that finds first negative number, if there are none 
> a value of 365
> will be output
> firstneg = function(x)if(any(x<0))min(which(x<0)) else 335
> 
> #apply the function to every column in Matrix
> FN <- apply(RM,2,firstneg)
> 
> #problem: column values are off by 30 (31 day running mean)
> FN <- matrix(FN) 
> add = function(x)x=x+30
> FN <- apply(FN,2,add)
> 
> #function that finds last negative number, if there are none 
> a value of 365
> will be output
> lastneg = function(x)if(any(x<0))max(which(x<0)) else 335
> 
> #apply the function to every column in Matrix
> LN <- apply(RM,2,lastneg)
> 
> #problem: column values are off by 30
> LN <- matrix(LN) 
> LN <- apply(LN,2,add)
> RMT <- t(RM)
> 
> #Create and fill an empty matrix with zeros
> sum <- matrix(nrow=R,ncol=1)
> for(i in 1:R){sum[i] <- 0}
> 
> #******This is the loop that I would like to speed up************
> 
> #sum values between FN and LN, excluding positive values
> for(i in 1:R){
> for(j in FN[i]:LN[i]){
> if(Temp[i,j]<0)
> sum[i] <- sum[i] + sum(Temp[i,j])}}
> 
Using sum for a matrix name was very confusing to my eyes trying to
differentiate square brackets from parentheses.  You may have even confused
yourself.  However, does this get you close to what you want?

#sum values between FN and LN, excluding positive values
for(i in 1:R){
  sum[i] <- sum(Temp[i,Temp[I,FN[i]:LN[i]]>0]) }
 
Hope this helpful,

Dan

Daniel J. Nordlund
Washington State Department of Social and Health Services
Planning, Performance, and Accountability
Research and Data Analysis Division
Olympia, WA  98504-5204