[R] sweep() and recycling

[Heather Turner]

Dr H Turner
Research Assistant
Dept. of Statistics
The University of Warwick
Coventry
CV4 7AL

Tel: 024 76575870

>>> Robin Hankin <r.hankin at noc.soton.ac.uk> 06/20/05 10:32am >>>
Hi


I had a hard-to-find bug in some of my code the other day, which I 
eventually
traced to my misusing of sweep().

I would expect sweep() to give
me a warning if the elements don't recycle nicely, but

  X <- matrix(1:36,6,6)
  sweep(X,1,1:5,"+")
      [,1] [,2] [,3] [,4] [,5] [,6]
[1,]    2    9   16   23   30   32
[2,]    4   11   18   25   27   34
[3,]    6   13   20   22   29   36
[4,]    8   15   17   24   31   38
[5,]   10   12   19   26   33   40
[6,]    7   14   21   28   35   37

gives no warning, even though (5,36)=1.

Also,


sweep(X,1,c(1,1000),"+")

and

sweep(X,2,c(1,1000),"+")


behave as expected,   But


sweep(X,1,1:5,"+")

and

  sweep(X,2,1:5,"+")



are identical!

...which is also expected since the matrix is square. Perhaps the
following will help you see why:
> matrix(1:5,6,6)
     [,1] [,2] [,3] [,4] [,5] [,6]
[1,]    1    2    3    4    5    1
[2,]    2    3    4    5    1    2
[3,]    3    4    5    1    2    3
[4,]    4    5    1    2    3    4
[5,]    5    1    2    3    4    5
[6,]    1    2    3    4    5    1
Warning message:
data length [5] is not a sub-multiple or multiple of the number of rows
[6] in matrix 
> matrix(1:5,6,6, byrow = TRUE)
     [,1] [,2] [,3] [,4] [,5] [,6]
[1,]    1    2    3    4    5    1
[2,]    2    3    4    5    1    2
[3,]    3    4    5    1    2    3
[4,]    4    5    1    2    3    4
[5,]    5    1    2    3    4    5
[6,]    1    2    3    4    5    1
Warning message:
data length [5] is not a sub-multiple or multiple of the number of rows
[6] in matrix 
i.e. recycling a vector of length 5 down the columns gives the same
result as recycling the same vector along the rows.

I agree with your main point however, it would be useful if 'sweep'
would give a warning when the length of the vector of statistics to be
swept out was not a sub-multiple or multiple of the corresponding array
dimension.

Heather