R help - Jan 2007 - Speeding things up

Hi,

is it possible to do this operation faster? I am going over 35k data
entries and this takes quite some time.

	for(cnt in 2:length(sdata$date))
	{

		if(sdata$value[cnt] < sdata$value[cnt - 1])	{
			sdata$ddtd[cnt] <- sdata$ddtd[cnt - 1] + sdata$value[cnt - 1] -
sdata$value[cnt]
		}
		else sdata$ddtd[cnt] <- 0
	
	}
	return(sdata)

Thank you,
Benjamin

First, since you only update the 'ddtd' conditioned on 'value',
you
should be able to vectorize removing the loop.  I let you figure out
how to do that yourself.

Second, you apply the "$" operator multiple times in the loop that
will definitely add some overhead.  It should be faster to extract
'value' and 'ddtd' and work with those instead.

/Henrik

On 1/8/07, Benjamin Dickgiesser <dickgiesser at gmail.com>
wrote:
> Hi,
>
> is it possible to do this operation faster? I am going over 35k data
> entries and this takes quite some time.
>
>         for(cnt in 2:length(sdata$date))
>         {
>
>                 if(sdata$value[cnt] < sdata$value[cnt - 1])     {
>                         sdata$ddtd[cnt] <- sdata$ddtd[cnt - 1] +
sdata$value[cnt - 1] -
> sdata$value[cnt]
>                 }
>                 else sdata$ddtd[cnt] <- 0
>
>         }
>         return(sdata)
>
> Thank you,
> Benjamin
>
> ______________________________________________
> R-help at stat.math.ethz.ch 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.
>

Hi Benjamin!


##TRY THIS: THIS MAYBE MUCH FASTER, BECAUSE ONLY WORK WITH THE IMPORTANAT
ROWS - HOPE NO ERROR IN IT:D

puffer1 <- as.matrix(sdata$value)
puffer2 <- rbind(as.matrix(puffer1[2:nrow(puffer1),1]),0)

speedy <- puffer1 > puffer2
speedy <- (as.matrix(which(puffer1==TRUE)))+1
sdata$ddtd[]=0

for(cnt in 1:nrow(speedy))
       {

sdata$ddtd[speedy[cnt,1]] <- sdata$ddtd[(speedy[cnt,1]) - 1] +
sdata$value[(speedy[cnt,1]) - 1] -sdata$value[speedy[cnt,1]]
       }
       return(sdata)



#Zoltan



2007/1/8, Benjamin Dickgiesser
<dickgiesser@gmail.com>:
>
> Hi,
>
> is it possible to do this operation faster? I am going over 35k data
> entries and this takes quite some time.
>
>        for(cnt in 2:length(sdata$date))
>        {
>
>                if(sdata$value[cnt] < sdata$value[cnt - 1])     {
>                        sdata$ddtd[cnt] <- sdata$ddtd[cnt - 1] +
> sdata$value[cnt - 1] -
> sdata$value[cnt]
>                }
>                else sdata$ddtd[cnt] <- 0
>
>        }
>        return(sdata)
>
> Thank you,
> Benjamin
>
> ______________________________________________
> R-help@stat.math.ethz.ch 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.
>
	[[alternative HTML version deleted]]

Thanks for your answers, your help is highly appreciated!

On 1/8/07, Zoltan Kmetty <zkmetty at gmail.com>
wrote:
> Hi Benjamin!
>
>
> ##TRY THIS: THIS MAYBE MUCH FASTER, BECAUSE ONLY WORK WITH THE IMPORTANAT
> ROWS - HOPE NO ERROR IN IT:D
>
> puffer1 <- as.matrix(sdata$value)
> puffer2 <- rbind(as.matrix(puffer1[2:nrow(puffer1),1]),0)
>
> speedy <- puffer1 > puffer2
> speedy <- (as.matrix(which(puffer1==TRUE)))+1
> sdata$ddtd[]=0
>
> for(cnt in 1:nrow(speedy))
>        {
>
> sdata$ddtd[speedy[cnt,1]] <- sdata$ddtd[(speedy[cnt,1]) - 1] +
> sdata$value[(speedy[cnt,1]) - 1] -sdata$value[speedy[cnt,1]]
>        }
>        return(sdata)
>
>
>
>
> #Zoltan
>
>
>
> 2007/1/8, Benjamin Dickgiesser <dickgiesser at gmail.com>:
> >
> > Hi,
> >
> > is it possible to do this operation faster? I am going over 35k data
> > entries and this takes quite some time.
> >
> >        for(cnt in 2:length(sdata$date))
> >        {
> >
> >                if(sdata$value[cnt] < sdata$value[cnt - 1])     {
> >                        sdata$ddtd[cnt] <- sdata$ddtd[cnt - 1] +
> sdata$value[cnt - 1] -
> > sdata$value[cnt]
> >                }
> >                else sdata$ddtd[cnt] <- 0
> >
> >        }
> >        return(sdata)
> >
> > Thank you,
> > Benjamin
> >
> > ______________________________________________
> > R-help at stat.math.ethz.ch 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.
> >
>
>