R help - Jan 2005 - avoiding loops

Hi again,

thanks a lot for the quick answer. I just forgot the comma, always 
these stupid mistakes...

Anyways, as I said before, I have two data.frames containing about 1000 
rows and I would like to avoid looping through all of them...

In each data.frame are coordinates (x,y,z), so every row is giving the 
information on one single point.
I would like to calculate the distance from each point in the one frame 
to every point in the second...
But how to do this without loops, how to do it quickly? I would 
appreciate it very much, if someone would introduce me to the high art 
of R - avoiding loops :-).

Cheers, Dax.

Doesn't this work?
> d1 <- data.frame(x=rnorm(100), y=rnorm(100), z=rnorm(100))
> d2 <- data.frame(x=rnorm(100), y=rnorm(100), z=rnorm(100))
> dist.12 <- sqrt(rowSums((d1 - d2)^2))
Andy
> From: dax42
> 
> Hi again,
> 
> thanks a lot for the quick answer. I just forgot the comma, always 
> these stupid mistakes...
> 
> Anyways, as I said before, I have two data.frames containing 
> about 1000 
> rows and I would like to avoid looping through all of them...
> 
> In each data.frame are coordinates (x,y,z), so every row is 
> giving the 
> information on one single point.
> I would like to calculate the distance from each point in the 
> one frame 
> to every point in the second...
> But how to do this without loops, how to do it quickly? I would 
> appreciate it very much, if someone would introduce me to the 
> high art 
> of R - avoiding loops :-).
> 
> Cheers, Dax.
> 
> ______________________________________________
> 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
> 
>

I didn't read the question carefully.  You were asking all pairwise
distances between the two sets, right?

This isn't the most efficient way to do it, but it seems OK for problem of
the size you stated:
> pairdist <- function(m1, m2) {
+     idx1 <- rep(1:nrow(m1), each=nrow(m2))
+     idx2 <- rep(1:nrow(m2), nrow(m1))
+     d <- sqrt(rowSums((m1[idx1,] - m2[idx2,])^2))
+     dim(d) <- c(nrow(m1), nrow(m2))
+     d
+ }
> d1 <- matrix(rnorm(3e3), 1e3, 3)
> d2 <- matrix(rnorm(3e3), 1e3, 3)
> system.time(dm <- pairdist(d1, d2), gcFirst=TRUE)
[1] 0.73 0.13 0.92   NA   NA
> str(dm)
 num [1:1000, 1:1000] 2.44 2.21 2.45 1.58 3.58 ...

Andy

> From: Liaw, Andy
> 
> Doesn't this work?
> 
> > d1 <- data.frame(x=rnorm(100), y=rnorm(100), z=rnorm(100))
> > d2 <- data.frame(x=rnorm(100), y=rnorm(100), z=rnorm(100))
> > dist.12 <- sqrt(rowSums((d1 - d2)^2))
> 
> Andy
> 
> > From: dax42
> > 
> > Hi again,
> > 
> > thanks a lot for the quick answer. I just forgot the comma, always 
> > these stupid mistakes...
> > 
> > Anyways, as I said before, I have two data.frames containing 
> > about 1000 
> > rows and I would like to avoid looping through all of them...
> > 
> > In each data.frame are coordinates (x,y,z), so every row is 
> > giving the 
> > information on one single point.
> > I would like to calculate the distance from each point in the 
> > one frame 
> > to every point in the second...
> > But how to do this without loops, how to do it quickly? I would 
> > appreciate it very much, if someone would introduce me to the 
> > high art 
> > of R - avoiding loops :-).
> > 
> > Cheers, Dax.
> > 
> > ______________________________________________
> > 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
> > 
> >
> 
> ______________________________________________
> 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
> 
> 
> --------------------------------------------------------------
> ----------------
> Notice:  This e-mail message, together with any attachments, 
> contains information of Merck & Co., Inc. (One Merck Drive, 
> Whitehouse Station, New Jersey, USA 08889), and/or its 
> affiliates (which may be known outside the United States as 
> Merck Frosst, Merck Sharp & Dohme or MSD and in Japan, as 
> Banyu) that may be confidential, proprietary copyrighted 
> and/or legally privileged. It is intended solely for the use 
> of the individual or entity named on this message.  If you 
> are not the intended recipient, and have received this 
> message in error, please notify us immediately by reply 
> e-mail and then delete it from your system.
> --------------------------------------------------------------
> ----------------
>

From:   Frank Samuelson <fws4 at cdrh.fda.gov>
> Along the lines of this thread,
> is there a general apply type function that
> allows me to take one vector at a time from a matrix (or
> row from a data frame) and another vector from another matrix
> and apply them on a general function? Sort of a multidimensional
> mapply, or an 'inner' routine where you can define the
> function (like you can for 'outer'.)
> 
> I've found that I do this regularly in my coding and I usually
> end up using indices (which are pretty fast in R):
> inner<-function(a,b,FUN=function(x,y) sum(x*y))
> sapply(1:nrow(a),function(i) FUN(a[i,],b[,i]))
> 
> Is there existing routine that does this?
> 
inner <- function(a,b,f) {
	f <- match.fun(f)
	apply(b,2,function(x)apply(a,1,function(y)f(x,y)))
}

# e.g. 
inner(matrix(1:4,2), matrix(4:1,2), crossprod)

matrix(1:4,2) %*% matrix(4:1,2) # same