Hi all,
I need to aggregate some matrix data (1440x720) to a lower dimension (720x360)
for lots of years and variables
I can do double for loop, but that will be slow. Anybody know a quicker way?
here an example with a smaller matrix size:
tst=matrix(1:(8*4),ncol=8,nrow=4)
tst_2x2=matrix(NA,ncol=4,nrow=2)
nx=2
ny=2
for(ilon in seq(1,8,nx)) {
for (ilat in seq(1,4,ny)) {
ilon_2x2=1+(ilon-1)/nx
ilat_2x2=1+(ilat-1)/ny
tst_2x2[ilat_2x2,ilon_2x2] = mean(tst[ilat+0:1,ilon+0:1])
}
}
tst
tst_2x2
> tst
[,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]
[1,] 1 5 9 13 17 21 25 29
[2,] 2 6 10 14 18 22 26 30
[3,] 3 7 11 15 19 23 27 31
[4,] 4 8 12 16 20 24 28 32
> tst_2x2
[,1] [,2] [,3] [,4]
[1,] 3.5 11.5 19.5 27.5
[2,] 5.5 13.5 21.5 29.5
I though a cast to 3d-array might do the trick and apply over the new dimension,
but that does not work, since it casts the data along the
row.> matrix(apply(array(tst,dim=c(nx,ny,8)),3,mean),nrow=nrow(tst)/ny)
[,1] [,2] [,3] [,4]
[1,] 2.5 10.5 18.5 26.5
[2,] 6.5 14.5 22.5 30.5
cheers
Peter
This should be faster. It uses apply() across the blocks.> ilon <- seq(1,8,nx) > ilat <- seq(1,4,ny) > cells <- as.matrix(expand.grid(ilat, ilon)) > blocks <- apply(cells, 1, function(x) tst[x[1]:(x[1]+1), x[2]:(x[2]+1)]) > block.means <- colMeans(blocks) > tst_2x2 <- matrix(block.means, 2, 4) > tst_2x2[,1] [,2] [,3] [,4] [1,] 3.5 11.5 19.5 27.5 [2,] 5.5 13.5 21.5 29.5 ------------------------------------- David L Carlson Department of Anthropology Texas A&M University College Station, TX 77840-4352 -----Original Message----- From: R-help [mailto:r-help-bounces at r-poject.org] On Behalf Of Anthoni, Peter (IMK) Sent: Wednesday, July 27, 2016 6:14 AM To: r-help at r-project.org Subject: [R] Aggregate matrix in a 2 by 2 manor Hi all, I need to aggregate some matrix data (1440x720) to a lower dimension (720x360) for lots of years and variables I can do double for loop, but that will be slow. Anybody know a quicker way? here an example with a smaller matrix size: tst=matrix(1:(8*4),ncol=8,nrow=4) tst_2x2=matrix(NA,ncol=4,nrow=2) nx=2 ny=2 for(ilon in seq(1,8,nx)) { for (ilat in seq(1,4,ny)) { ilon_2x2=1+(ilon-1)/nx ilat_2x2=1+(ilat-1)/ny tst_2x2[ilat_2x2,ilon_2x2] = mean(tst[ilat+0:1,ilon+0:1]) } } tst tst_2x2> tst[,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [1,] 1 5 9 13 17 21 25 29 [2,] 2 6 10 14 18 22 26 30 [3,] 3 7 11 15 19 23 27 31 [4,] 4 8 12 16 20 24 28 32> tst_2x2[,1] [,2] [,3] [,4] [1,] 3.5 11.5 19.5 27.5 [2,] 5.5 13.5 21.5 29.5 I though a cast to 3d-array might do the trick and apply over the new dimension, but that does not work, since it casts the data along the row.> matrix(apply(array(tst,dim=c(nx,ny,8)),3,mean),nrow=nrow(tst)/ny)[,1] [,2] [,3] [,4] [1,] 2.5 10.5 18.5 26.5 [2,] 6.5 14.5 22.5 30.5 cheers Peter ______________________________________________ R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.
An alternative (more compact, not necessarily faster, because apply is still a
for loop inside):
f <- function( m, nx, ny ) {
# redefine the dimensions of my
a <- array( m
, dim = c( ny
, nrow( m ) %/% ny
, ncol( m ) %/% nx )
)
# apply mean over dim 1
apply( a, c( 2, 3 ), FUN=mean )
}
f( tst, nx, ny )
--
Sent from my phone. Please excuse my brevity.
On July 27, 2016 9:08:32 AM PDT, David L Carlson <dcarlson at tamu.edu>
wrote:>This should be faster. It uses apply() across the blocks.
>
>> ilon <- seq(1,8,nx)
>> ilat <- seq(1,4,ny)
>> cells <- as.matrix(expand.grid(ilat, ilon))
>> blocks <- apply(cells, 1, function(x) tst[x[1]:(x[1]+1),
>x[2]:(x[2]+1)])
>> block.means <- colMeans(blocks)
>> tst_2x2 <- matrix(block.means, 2, 4)
>> tst_2x2
> [,1] [,2] [,3] [,4]
>[1,] 3.5 11.5 19.5 27.5
>[2,] 5.5 13.5 21.5 29.5
>
>-------------------------------------
>David L Carlson
>Department of Anthropology
>Texas A&M University
>College Station, TX 77840-4352
>
>
>
>-----Original Message-----
>From: R-help [mailto:r-help-bounces at r-poject.org] On Behalf Of Anthoni,
>Peter (IMK)
>Sent: Wednesday, July 27, 2016 6:14 AM
>To: r-help at r-project.org
>Subject: [R] Aggregate matrix in a 2 by 2 manor
>
>Hi all,
>
>I need to aggregate some matrix data (1440x720) to a lower dimension
>(720x360) for lots of years and variables
>
>I can do double for loop, but that will be slow. Anybody know a quicker
>way?
>
>here an example with a smaller matrix size:
>
>tst=matrix(1:(8*4),ncol=8,nrow=4)
>tst_2x2=matrix(NA,ncol=4,nrow=2)
>nx=2
>ny=2
>for(ilon in seq(1,8,nx)) {
> for (ilat in seq(1,4,ny)) {
> ilon_2x2=1+(ilon-1)/nx
> ilat_2x2=1+(ilat-1)/ny
> tst_2x2[ilat_2x2,ilon_2x2] = mean(tst[ilat+0:1,ilon+0:1])
> }
>}
>
>tst
>tst_2x2
>
>> tst
> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]
>[1,] 1 5 9 13 17 21 25 29
>[2,] 2 6 10 14 18 22 26 30
>[3,] 3 7 11 15 19 23 27 31
>[4,] 4 8 12 16 20 24 28 32
>
>> tst_2x2
> [,1] [,2] [,3] [,4]
>[1,] 3.5 11.5 19.5 27.5
>[2,] 5.5 13.5 21.5 29.5
>
>
>I though a cast to 3d-array might do the trick and apply over the new
>dimension, but that does not work, since it casts the data along the
>row.
>> matrix(apply(array(tst,dim=c(nx,ny,8)),3,mean),nrow=nrow(tst)/ny)
> [,1] [,2] [,3] [,4]
>[1,] 2.5 10.5 18.5 26.5
>[2,] 6.5 14.5 22.5 30.5
>
>
>cheers
>Peter
>
>______________________________________________
>R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>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 at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>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.