Prof Brian Ripley wrote:
On Thu, 21 Apr 2005, Ernesto Jardim wrote:
I'm using colSums and rowSums to sum the first dimensions of arrays.
It works ok but the resulting object is different. See
a3d <- array(rnorm(120, mean=2), dim=c(20,6,1))
dim(colSums(a3d))
[1] 6 1
dim(rowSums(a3d))
NULL
class(colSums(a3d))
[1] "matrix"
class(rowSums(a3d))
[1] "numeric"
I was expecting rowSums to preserve the array class and the relevant
dimensions (1,20).
The main problem is with arrays where the third dimension (or higher)
is > 1. colSums preserve the array but rowSums concatenate the results:
No, it acts as documented.
a3d <- array(rnorm(120, mean=2), dim=c(20,3,2))
rowSums(a3d)
[1] 8.894178 11.932361 15.231601 12.374629 11.823671 10.564709 9.065166
[8] 13.900264 13.331756 9.351242 11.989821 7.643745 9.923288 8.169997
[15] 12.124624 16.711742 11.414150 15.221880 12.053734 13.368988
colSums(a3d)
[,1] [,2]
[1,] 44.80941 29.49216
[2,] 42.18339 39.81121
[3,] 39.90528 38.89010
Is this on purpose ?
Yes, and documented. rowSums(a3d) is a vector of length 20, by
dims: Which dimensions are regarded as "rows" or "columns" to sum
over. For 'row*', the sum or mean is over dimensions
'dims+1, ...'; for 'col*' it is over dimensions '1:dims'.
whereas colSums(a3d) is an array of dims c(6, 1).
Don't be confused by the class: there is no S3 class here.
Ok,
Thanks for the enlightment.
EJ
______________________________________________
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